- Overview of Secure Connectivity
-
- Implementing and Managing PKI Features Roadmap
- Overview of PKI
- Deploying RSA Keys Within a PKI
- Configuring Authorization and Revocation of Certificates in a PKI
- Configuring Certificate Enrollment for a PKI
- Storing PKI Credentials
- Source Interface Selection for Outgoing Traffic with Certificate Authority
- Cisco Group Encrypted Transport VPN
- Finding Feature Information
- Contents
- Prerequisites for Cisco Group Encrypted Transport VPN
- Restrictions for Cisco Group Encrypted Transport VPN
- Information About Cisco Group Encrypted Transport VPN
- Cisco Group Encrypted Transport VPN Overview
- Cisco Group Encrypted Transport VPN Architecture
- Cisco Group Encrypted Transport VPN Features
- Rekeying
- Group Member Access Control List
- Fail-Close Mode
- Time-Based Antireplay
- Cooperative Key Server
- Change Key Server Role
- Receive Only SA
- Passive SA
- Enhanced Solutions Manageability
- Support with VRF-Lite Interfaces
- GET VPN VRF-Aware GDOI on GM
- Authentication Policy for GM Registration
- Rekey Functionality in Protocol Independent Multicast-Sparse Mode
- Configuring a Key Server
- Prerequisites
- Configuring RSA Keys to Sign Rekey Messages
- Configuring the Group ID, Server Type, and SA Type
- Configuring the Rekey
- Configuring Group Member ACLs
- Configuring an IPsec Lifetime Timer
- Configuring an ISAKMP Lifetime Timer
- Configuring the IPsec SA
- Configuring Time-Based Antireplay for a GDOI Group
- Configuring Passive SA
- Resetting the Role of the Key Server
- Configuring a Group Member
- Configuring the Group Name, ID, Key Server IP Address, and Group Member Registration
- Creating a Crypto Map Entry
- Applying the Crypto Map to an Interface to Which the Traffic Must Be Encrypted
- Activating Fail-Close Mode
- Configuring Ciphers or Hash Algorithms for KEK
- Configuring Acceptable Transform Sets for TEK
- Configuring GET VPN GM Authorization
- Verifying and Troubleshooting Cisco Group Encrypted Transport VPN Configurations
- Verifying Active Group Members on a Key Server
- Verifying Rekey-Related Statistics
- Verifying IPsec SAs That Were Created by GDOI on a Group Member
- Verifying IPsec SAs That Were Created by GDOI on a Key Server
- Verifying Cooperative Key Server States and Statistics
- Verifying Antireplay Pseudotime-Related Statistics
- Verifying the Fail-Close Mode Status of a Crypto Map
Cisco Group Encrypted Transport VPN
Last Updated: October 2, 2011
Cisco Group Encrypted Transport Virtual Private Network (GET VPN) is a set of features that are necessary to secure IP multicast group traffic or unicast traffic over a private WAN that originates on or flows through a Cisco IOS device. GET VPN combines the keying protocol Group Domain of Interpretation (GDOI) with IP security (IPsec) encryption to provide users with an efficient method to secure IP multicast traffic or unicast traffic. GET VPN enables the router to apply encryption to nontunneled (that is, "native") IP multicast and unicast packets and eliminates the requirement to configure tunnels to protect multicast and unicast traffic.
This document describes how to configure, verify, and troubleshoot Cisco GET VPN.
Cisco Group Encrypted Transport VPN provides the following benefits:
•Provides data security and transport authentication, helping to meet security compliance and internal regulation by encrypting all WAN traffic
•Enables high-scale network meshes and eliminates complex peer-to-peer key management with group encryption keys
•For Multiprotocol Label Switching (MPLS) networks, maintains network intelligence such as full-mesh connectivity, natural routing path, and quality of service (QoS)
•Grants easy membership control with a centralized key server
•Helps ensure low latency and jitter by enabling full-time, direct communications between sites, without requiring transport through a central hub
•Reduces traffic loads on customer premises equipment (CPE) and provider-edge (PE) encryption devices by using the core network for replication of multicast traffic, avoiding packet replication at each individual peer site
Finding Feature Information
Your software release may not support all the features documented in this module. For the latest feature information and caveats, see the release notes for your platform and software release. To find information about the features documented in this module, and to see a list of the releases in which each feature is supported, see the "Feature Information for Cisco Group Encrypted Transport VPN" section.
Use Cisco Feature Navigator to find information about platform support and Cisco software image support. To access Cisco Feature Navigator, go to http://www.cisco.com/go/cfn. An account on Cisco.com is not required.
Contents
•Prerequisites for Cisco Group Encrypted Transport VPN
•Restrictions for Cisco Group Encrypted Transport VPN
•Information About Cisco Group Encrypted Transport VPN
•How to Configure Cisco Group Encrypted Transport VPN
•Configuration Examples for Cisco Group Encrypted Transport VPN
•Feature Information for Cisco Group Encrypted Transport VPN
•Appendix I: System Logging Messages
Prerequisites for Cisco Group Encrypted Transport VPN
•You must be using Cisco IOS XE Release 2.3 or later.
•You should be knowledgeable about IPsec and Internet Key Exchange (IKE).
•You should know how to configure multicast and unicast routing on a Cisco IOS XE global router.
•When the IKE policy is configured, the IKE lifetime should be set to the minimum of 5 minutes so that unnecessary resources are not wasted on the maintenance of the IKE security association (SA). After the registration IKE SA is established, the registration SAs no longer have to be maintained because the rekey SA has been created and will be used to accept future rekeys.
Restrictions for Cisco Group Encrypted Transport VPN
•If you are encrypting high packet rates for counter-based antireplay, ensure that you do not make the lifetime too long or it can take several hours for the sequence number to wrap. For example, if the packet rate is 100 kilopackets per second, the lifetime should be configured as fewer than 11.93 hours so that the SA is used before the sequence number wraps.
•For unicast traffic and counter-based antireplay, the sequence numbers may be out of sync between the group members if one of the group members goes down and comes back up. For example: There is traffic from group member 1 to group member 2, and the last sequence number is n. Group member 1 goes down and comes back up. The sequence number of the SA at group member 1 now starts with 1, but group member 2 is expecting continuation from the previous sequence number (n + 1). This situation causes subsequent traffic from group member 1 to be dropped until the sequence number on group member 1 reaches n or the next rekey.
•When you configure transport mode traffic selectors, it is possible to have transport mode SAs. SAs occur when the packet size exceeds the MTU, and the packet cannot be forwarded.
•Transport mode should be used only for Group Encrypted Transport VPN Mode (GM) to GM traffic.
•If you are overriding the don't fragment bit (df-bit) setting in the IP header of encapsulated packets, you must configure the override commands in global configuration mode. GET VPN does not honor the interface configuration. This restriction is limited only to GET VPN. IPsec accepts both global configuration- and interface-specific override commands.
•Counter-based antireplay is not recommended and works only if there are two group members in a group.
•Because Path MTU Discovery (PMTUD) does not work for GET VPN, there is a possibility that encapsulated packets could be dropped when the df-bit is set and the MTU of an intermediate link is less than the size of the encapsulated packet. In such an event, the router that drops the packet sends a notification to the source IP address on the packet, indicating that the packet has been dropped because the router could not fragment the packet due to the df-bit setting. In GET VPN, this message goes past the encapsulating endpoint directly to the source of the data due to the header preservation feature of GET VPN. Thus, the encapsulating router never knows that it has to fragment the packet to a smaller size before setting the df-bit after encapsulation. It continues to set the df-bit on the packets and they continue to be dropped at the intermediate router. (This is known as black-holing the traffic.)
•Key servers cannot be configured using Cisco IOS XE images. They must be configured using Cisco IOS T-based or mainline-based images.
•Because of crypto engine optimization, the time-based antireplay (TBAR) overhead is 16 bytes.
•The GDOI protocol does not allow port ranges on the crypto access control list (ACL) policy on a key server. Therefore, you must specify any deny ACL statements containing port ranges locally on each GM (where they will be merged with the policy that is received via GDOI).
Information About Cisco Group Encrypted Transport VPN
•Cisco Group Encrypted Transport VPN Overview
•Cisco Group Encrypted Transport VPN Architecture
•Cisco Group Encrypted Transport VPN Features
Cisco Group Encrypted Transport VPN Overview
Networked applications such as voice and video increase the need for instantaneous, branch-interconnected, and QoS-enabled WANs. The distributed nature of these applications results in increased demands for scale. At the same time, enterprise WAN technologies force businesses to trade off between QoS-enabled branch interconnectivity and transport security. As network security risks increase and regulatory compliance becomes essential, GET VPN, a next-generation WAN encryption technology, eliminates the need to compromise between network intelligence and data privacy.
With GET, Cisco provides tunnelless VPN, which eliminates the need for tunnels. Meshed networks, by removing the need for point-to-point tunnels, can scale higher while maintaining network-intelligence features critical to voice and video quality. GET is a standards-based security model that is based on the concept of "trusted" group members. Trusted member routers use a common security methodology that is independent of any point-to-point IPsec tunnel relationship. Also, "any-any" networks, by using trusted groups instead of point-to-point tunnels, can scale higher while maintaining network-intelligence features (such as QoS, routing, and multicast), which are critical to voice and video quality.
GET-based networks can be used in a variety of WAN environments, including IP and MPLS. MPLS VPNs that use this encryption technology are highly scalable, manageable, and cost-effective, and they meet government-mandated encryption requirements. The flexible nature of GET allows security-conscious enterprises either to manage their own network security over a service provider WAN service or to offload encryption services to their providers. GET simplifies securing large Layer 2 or MPLS networks that require partial or full-mesh connectivity.
Cisco Group Encrypted Transport VPN Architecture
GET VPN encompasses Multicast Rekeying, a way to enable encryption for "native" multicast packets, and unicast rekeying over a private WAN. Multicast Rekeying and GET VPN is based on GDOI as defined in Internet Engineering Task Force (IETF) RFC 3547. In addition, there are similarities to IPsec in the area of header preservation and SA lookup. Dynamic distribution of IPsec SAs has been added, and tunnel overlay properties of IPsec have been removed. Figure 1 further illustrates the GET VPN concepts and relationships.
Figure 1 GET VPN Concepts and Relationships
This section has the following subsections:
•Key Distribution: Group Domain of Interpretation
Key Distribution: Group Domain of Interpretation
GDOI
GDOI is defined as the Internet Security Association Key Management Protocol (ISAKMP) Domain of Interpretation (DOI) for group key management. In a group management model, the GDOI protocol operates between a group member and a group controller or key server (GCKS), which establishes SAs among authorized group members. The ISAKMP defines two phases of negotiation. GDOI is protected by a Phase 1 ISAKMP security association. The Phase 2 exchange is defined in IETF RFC 3547. The topology shown in Figure 2 and the corresponding explanation show how this protocol works.
Group Member
The group member registers with the key server to get the IPsec SA or SAs that are necessary to communicate with the group. The group member provides the group ID to the key server to get the respective policy and keys for this group. These keys are refreshed periodically, and before the current IPsec SAs expire, so that there is no loss of traffic.
The output of the show crypto isakmp sa detail command will show the security association (SA) Authentication as "rsig" because the RSA signature is used for key encryption key (KEK) rekey authentication in GET VPN.
Key Server
The responsibilities of the key server include maintaining the policy and creating and maintaining the keys for the group. When a group member registers, the key server downloads this policy and the keys to the group member. The key server also rekeys the group before existing keys expire.
Note Key servers are not supported on the Cisco ASR 1000 series routers. They must be configured using Cisco IOS T-based or mainline-based images.
The key server has two responsibilities: servicing registration requests and sending rekeys. A group member can register at any time and receive the most current policy and keys. When a group member registers with the key server, the key server verifies the group ID that the group member is attempting to join. If this ID is a valid group ID, the key server sends the SA policy to the group member. After the group member acknowledges that it can handle the downloaded policy, the key server downloads the respective keys.
There are two types of keys that the key server can download: the key encryption key (KEK) and the traffic encryption key (TEK). The TEK becomes the IPsec SA with which the group members within the same group communicate. The KEK encrypts the rekey message.
The GDOI server sends out rekey messages if an impending IPsec SA expiration occurs or if the policy has changed on the key server (using the command-line interface [CLI]). A rekey can also happen if the KEK timer has expired, and the key server sends out a KEK rekey. The rekey messages may also be retransmitted periodically to account for possible packet loss. Packet loss can occur because rekey messages are sent without the use of any reliable transport. If the rekey mechanism is multicast, there is no efficient feedback mechanism by which receivers can indicate that they did not receive a rekey message, so retransmission seeks to bring all receivers up to date. If the rekey mechanism is unicast, the receivers will send an acknowledgment message.
Figure 2 Protocol Flows That Are Necessary for Group Members
to Participate in a Group
The topology shows the protocol flows that are necessary for group members to participate in a group, which are as follows:
1. Group members register with the key server. The key server authenticates and authorizes the group members and downloads the IPsec policy and keys that are necessary for them to encrypt and decrypt IP multicast packets.
2. As needed, the key server "pushes" a rekey message to the group members. The rekey message contains a new IPsec policy and keys to use when old IPsec SAs expire. Rekey messages are sent in advance of the SA expiration time to ensure that valid group keys are always available.
3. The group members are authenticated by the key server and communicate with other authenticated group members that are in the same group using the IPsec SAs that the group members have received from the key server.
How Protocol Messages Work with Cisco IOS XE Software
Multicast Rekeying uses the GDOI protocol (IETF RFC 3547) to distribute the policy and keys for the group. The GDOI protocol is between a key server and a group member. The key server creates and maintains the policy and keys, and it downloads the policy and keys to the authenticated group members.
The GDOI protocol is protected by an ISAKMP Phase 1 exchange. The GDOI key server and the GDOI group member must have the same ISAKMP policy. This Phase 1 ISAKMP policy should be strong enough to protect the GDOI protocol that follows. The GDOI protocol is a four-message exchange that follows the Phase 1 ISAKMP policy. The Phase 1 ISAKMP exchange can occur in main mode or aggressive mode.
Figure 3 shows the ISAKMP Phase 1 exchange.
Figure 3 ISAKMP Phase 1 Exchange and GDOI Registration
The ISAKMP Phase 1 messages and the four GDOI protocol messages are referred to as the GDOI registration, and the entire exchange that is shown is a unicast exchange between the group member and the key server.
During the registration, if the rekey mechanism is multicast, the group member receives the address of the multicast group and registers with the multicast group that is required to receive the multicast rekeys.
The GDOI protocol uses User Datagram Protocol (UDP) port 848 (with Network Address Translation-Traversal (NAT-T), it floats to 4500).
IPsec
IPsec is a well-known RFC that defines an architecture to provide various security services for traffic at the IP layer. The components and how they fit together with each other and into the IP environment are described in IETF RFC 2401.
Communication Flow Between Key Servers and Group Members to Update IPsec SAs
Key servers and group members are the two components of the GET VPN architecture. The key server holds and supplies group authentication keys and IPsec SAs to the group members.
Group members provide encryption service to the interesting traffic (traffic that is worthy of being encrypted and secured by IPsec).
Communication among the key server and group members is encrypted and secured. GDOI supports the use of two keys: the TEK and the KEK. The TEK is downloaded by the key server to all the group members. The downloaded TEK is used by all the group members to communicate securely among each other. This key is essentially the group key that is shared by all the group members. The group policies and IPsec SAs are refreshed by the key server using periodic rekey messages to the group members. The KEK is also downloaded by the key server and is used by the group members to decrypt the incoming rekey messages from the key server.
The key server generates the group policy and IPsec SAs for the GDOI group. The information generated by the key server includes multiple TEK attributes, traffic encryption policy, lifetime, source and destination, a Security Parameter Index (SPI) ID that is associated with each TEK, and the rekey policy (one KEK).
Figure 4 illustrates the communication flow between group members and the key server. The key server, after receiving registration messages from a group member, generates the information that contains the group policy and new IPsec SAs. The new IPsec SA is then downloaded to the group member. The key server maintains a table that contains the IP address of each group member per group. When a group member registers, the key server adds its IP address in its associated group table, thus allowing the key server to monitor an active group member. A key server can support multiple groups. A group member can be part of multiple groups.
Figure 4 Communication Flow Between Group Members and the Key Server
IPsec and ISAKMP Timers
IPsec and ISAKMP SAs are maintained by the following timers:
•TEK lifetime—Determines the lifetime of the IPsec SA. Before the end of the TEK lifetime, the key server sends a rekey message, which includes a new TEK encryption key and transforms as well as the existing KEK encryption keys and transforms. The TEK lifetime is configured only on the key server, and the lifetime is "pushed down" to the group members using the GDOI protocol. The TEK lifetime value depends on the security policy of the network. If the set security-association lifetime command is not configured, the default value of 86,400 seconds takes effect. To configure a TEK lifetime, see the "Configuring an IPsec Lifetime Timer" section.
•KEK lifetime—Determines the lifetime of the GET VPN rekey SAs. Before the end of the lifetime, the key server sends a rekey message, which includes a new KEK encryption key and transforms and new TEK encryption keys and transforms. The KEK lifetime is configured only on the key server, and the lifetime is pushed down to group members dynamically using the GDOI protocol. The KEK lifetime value should be greater than the TEK lifetime value (it is recommended that the KEK lifetime value be at least three times greater than the TEK lifetime value). If the rekey lifetime command is not configured, the default value of 86,400 seconds takes effect. To configure a KEK lifetime, see the "Configuring a Multicast Rekey" section.
•ISAKMP SA lifetime—Defines how long each ISAKMP SA should exist before it expires. The ISAKMP SA lifetime is configured on a group member and on the key server. If the group members and key servers do not have a cooperative key server, the ISAKMP SA is not used after the group member registration. In this case (no cooperative key server), the ISAKMP SA can have a short lifetime (a minimum of 60 seconds). If there is a cooperative key server, all key servers must have long lifetimes to keep the ISAKMP SA "up" for cooperative key server communications. If the lifetime command is not configured, the default value of 86,400 seconds takes effect. To configure an ISAKMP SA lifetime, see the "Configuring an ISAKMP Lifetime Timer" section.
Address Preservation
The following section describes address preservation in GET VPN.
As shown in Figure 5, IPsec-protected data packets carry the original source and destination in the outer IP header rather than replacing them with tunnel endpoint addresses. This technique is known as IPsec Tunnel Mode with Address Preservation.
Figure 5 Header Preservation
Address preservation allows GET VPN to use the routing functionality present within the core network. Address preservation allows routing to deliver the packets to any customer-edge (CE) device in the network that advertises a route to the destination address. Any source and destination matching the policy for the group will be treated in a similar manner. In the situation where a link between IPsec peers is not available, address preservation also helps combat traffic "black-hole" situations.
Header preservation also maintains routing continuity throughout the enterprise address space and in the WAN. As a result, end host addresses of the campus are exposed in the WAN (for MPLS, this applies to the edge of the WAN). For this reason, GET VPN is applicable only when the WAN network acts as a "private" network (for example, in an MPLS network).
Secure Data Plane Multicast
The multicast sender uses the TEK that is obtained from the key server and encrypts the multicast data packet with header preservation before it switches out the packet. The replication of the multicast packet is carried out in the core on the basis of the (S, G) state that is retained in the multicast data packet. This process is illustrated in Figure 6.
Figure 6 Secure Data Plane Multicast Process
Secure Data Plane Unicast
The unicast sender uses the TEK that is obtained from the key server and encrypts the unicast data packet with header preservation before it switches out the packet to the destination. This process is illustrated in Figure 7.
Figure 7 Secure Data Plane Unicast Process
Cisco Group Encrypted Transport VPN Features
This section includes the following subsections:
•Group Member Access Control List
•Enhanced Solutions Manageability
•Support with VRF-Lite Interfaces
•Authentication Policy for GM Registration
•Rekey Functionality in Protocol Independent Multicast-Sparse Mode
Rekeying
Rekey messages are used to refresh IPsec SAs. When the IPsec SAs or the rekey SAs are about to expire, one single rekey message for a particular group is generated on the key server. No new IKE sessions are created for the rekey message distribution. The rekey messages are distributed by the key server over an existing IKE SA.
Rekeying can use multicast or unicast messages. GET VPN supports both unicast and multicast rekeying.
The following subsections give detailed rekeying information:
•Rekey Behavior After Policy Changes
•IPsec SA Usage on the Group Members
•Configuration Changes Can Trigger a Rekey By a Key Server
•Commands That Trigger a Rekey
Rekey Sequence-Number Check
The rekey sequence-number check between the key server and the group member is conducted as follows:
1. Antireplay in GROUPKEY-PUSH messages is restored as specified in RFC 3547.
–The group member drops any rekey message that has a sequence number lower than or equal to that of the last received rekey message.
–The group member accepts any rekey message that has a sequence number higher than that of the last received rekey message, no matter how large the difference.
2. The sequence number is reset to 1 at the first rekey message after the KEK rekey, not at the KEK rekey message itself.
Multicast Rekeying
Multicast rekeys are sent out using an efficient multicast rekey. Following a successful registration, the group member registers with a particular multicast group. All the group members that are registered to the group receives this multicast rekey. Multicast rekeys are sent out periodically on the basis of the configured lifetime on the key server. Multicast rekeys are also sent out if the IPsec or rekey policy is changed on the key server. Triggered by the configuration change, the rekey sends out the new updated policy to all the group members with an efficient multicast rekey.
The key server pushes the rekey time back as follows:
1. If the TEK timeout is 300 seconds:
tek_rekey_offset = 90 (because 300 < 900)
If retransmissions are configured, the rekey timer is moved back more.
For three retransmissions every 10 seconds: 3 x 10
So the rekey will actually happen at (300 - 90 - 30) = 180 seconds
2. If the TEK timeout is 3600 seconds:
tek_rekey_offset = 3600 x 10 percent = 360 seconds
If retransmissions are configured, the rekey timer is moved back more.
For three retransmissions every 10 seconds: 3 x 10
So the rekey will actually happen at (3600 - 360 - 30) = 3210 seconds
When a KEK expires and when the transport mode is multicast, a multicast KEK rekey is sent. When a multicast KEK rekey is sent, the group member replaces the old KEK with the new KEK. Because it is a multicast rekey, and the retransmissions are sent, the old KEK continues to be used for encryption. This situation occurs because the group member does not receive the new KEK rekey. Hence the group member that received the multicast KEK rekey does not have the old KEK, and hence it drops these retransmissions.
The group member that did not initially receive the KEK key now receives the KEK retransmission and replaces the old KEK with the new KEK and will drop the retransmissions that will follow. For example, if five retransmissions are configured and a multicast KEK rekey with sequence number 1 is received at group member 1, all the other retransmissions with sequence numbers 2 3 4 5 6 will be dropped at the group member because the group member does not have the old KEK.
If group member 2 does not get the KEK rekey with sequence number 1 and it receives the retransmission with sequence number 2, it will drop the other retransmissions 3, 4, 5, 6.
Unicast Rekeying and SAs
In a large unicast group, to alleviate latency issues, the key server generates rekey messages for only a small number of group members at a time. The key server is ensured that all group members receive the same rekey messages for the new SA before the expiration of the old SA. Also, in a unicast group, after receiving the rekey message from the key server, a group member sends an encrypted acknowledge (ACK) message to the key server using the keys that were received as part of the rekey message. When the key server receives this ACK message, it notes this receipt in its associated group table, which accomplishes the following:
•The key server keeps a current list of active group members.
•The key server sends rekey messages only to active members.
In addition, in a unicast group, the key server removes the group member from its active list and stops sending the rekey messages to that particular group member if the key server does not receive an ACK message for three consecutive rekeys. If no ACK message is received for three consecutive rekeys, the group member has to fully reregister with the key server after its current SA expires if the group member is still interested in receiving the rekey messages. The ejection of a nonresponsive group member is accomplished only when the key server is operating in the unicast rekey mode. The key server does not eject group members in the multicast rekey mode because group members cannot send ACK messages in that mode.
As in multicast rekeying, if retransmission is configured, each rekey will be retransmitted the configured number of times.
Rekey transport modes and authentication can be configured under a GDOI group.
If unicast rekey transport mode is not defined, multicast is applied by default.
If the TEK rekey is not received, the group member reregisters with the key server 60 seconds before the current IPsec SA expires. The key server has to send out the rekey before the group member reregistration occurs. If no retransmission is configured, the key server sends the rekey tek_rekey_offset before the SA expires. The tek_rekey_offset is calculated based on the configured rekey lifetime. If the TEK rekey lifetime is less than 900 seconds, the tek_rekey_offset is set to 90 seconds. If the TEK rekey lifetime is configured as more than 900 seconds, the tek_rekey_offset = (configured TEK rekey lifetime)/10. If retransmission is configured, the rekey occurs earlier than the tek_rekey_offset to let the last retransmission be sent 90 seconds before the SA expires.
The key server uses the formula in the following example to calculate when to start sending the rekey to all unicast group members. The unicast rekey process on the key server sends rekeys to unicast group members in groups of 50 within a loop. The time spent within this loop is estimated to be 5 seconds.
A key server rekeys group members in groups of 50, which equals two loops. For example, for 100 group members:
Number of rekey loops = (100 group members)/50 = 2 loops:
•Time required to rekey one loop (estimation) = 5 seconds
•Time to rekey 100 group members in two loops of 50: 2 x 5 seconds = 10 seconds
So the key server pushes the rekey time back as follows:
•If the TEK timeout is 300: 300 - 10 = 290
But the start has to be earlier than the TEK expiry (as in the multicast case):
•Because 300 < 900, tek_rekey_offset = 90
•So 90 seconds is subtracted from the actual TEK time: 290 - tek_rekey_offset = 200 seconds
If retransmissions are configured, the rekey timer is moved back more:
•For three retransmissions every 10 seconds: 200 - (3 x 10) = 170
•If the TEK timeout is 3600 seconds: 3600 - 10 = 3590
But the start has to be earlier than the TEK expiry (as in the multicast case):
•Because 3600 > 900, tek_rekey_offset = 3600 x 10 percent = 360
•So 360 seconds is subtracted from the actual TEK time: 3590 - tek_rekey_offset = 3230 seconds
If retransmissions are configured, the rekey timer is moved back more:
•For three retransmissions every 10 seconds: 3230 - (3 x 10) = 3200 seconds
The tek_rekey_offset formula applies to unicast and multicast rekeying.
Rekey Behavior After Policy Changes
Table 1 provides a list of rekey behavior based on the security policy changes.
Enter the following commands for the policy changes to take effect immediately:
•Use the clear crypto gdoi [group] command on the key server.
•Use the clear crypto gdoi [group] command on all the group members.
IPsec SA Usage on the Group Members
When a rekey is received and processed on a group member, the new IPsec SA (the SPI) is installed. There is a period of time when the old and the new IPsec SAs are used. After a certain specified interval, the old IPsec SA is deleted. This overlap ensures that all group members receive the current rekey and insert the new IPsec SAs. This behavior is independent of the transport method (multicast or unicast rekey transport) for the rekeys from the key server.
Approximately 30 seconds before the old SA expires, the group member starts to use the new SA in the outbound direction to encrypt the packet. Approximately 60 seconds before the old SA expires, if no new SA is received on the group member side via a rekey from the key server, the group member reregisters.
In Figure 8, time T2 is when the old SA expires. T1 is 30 seconds before T2, which is when the group member (GM) starts to use the new SA in the outbound direction. T0 is another 30 seconds before T2. If no new SA is received at T0, the group member has to reregister. T is another 30 seconds from T0. The key server should send a rekey at T.
Figure 8 IPsec SA Usage on a Group Member
Configuration Changes Can Trigger a Rekey By a Key Server
Configuration changes on a key server can trigger a rekey by the key server. Please refer to the following sample configuration as you read through the changes that will or will not cause a rekey that are described following the example.
crypto ipsec transform-set gdoi-p esp-3des esp-sha-hmac
!
crypto ipsec profile gdoi-p
set security-association lifetime seconds 900
set transform-set gdoi-p
!
crypto gdoi group diffint
identity number 3333
server local
rekey algorithm aes 128
rekey address ipv4 121
rekey lifetime seconds 3600
no rekey retransmit
rekey authentication mypubkey rsa mykeys
sa ipsec 1
profile gdoi-p
match address ipv4 120
replay counter window-size 3
The following configuration changes on the key server will trigger a rekey from the key server:
•Any change in the TEK configuration ("sa ipsec 1" in the example):
–If the ACL ("match address ipv4 120" in the above example) is changed. Any addition, deletion, or change in the ACL causes a rekey.
–If TEK replay is enabled or disabled on the key server, rekey is sent.
–Removal or addition of the IPsec profile in the TEK ("profile gdoi-p" in the example).
–Changing from multicast to unicast transport.
–Changing from unicast to multicast transport.
The following configuration changes on the key server will not trigger a rekey from the key server:
•Replay counter window size is changed under the TEK ("sa ipsec 1" in the example).
•Configuring or removing rekey retransmit.
•Removing or configuring the rekey ACL.
•Changing the TEK lifetime ("set security-association lifetime seconds 300" in the example) or changing the KEK lifetime ("rekey lifetime seconds 500" in the example).
•Adding, deleting, or changing the rekey algorithm ("rekey algorithm aes 128" in the example).
Commands That Trigger a Rekey
Table 2 is a comprehensive list of GET VPN command changes, and it shows which commands will or will not trigger a rekey. Commands are broken out based on the configuration mode in which they are entered. The table also shows when the commands take effect, regardless of whether they trigger a rekey.
When a timeout is caused by a pseudotime synchronization, the key server checks if either the KEK or the TEK timer is scheduled to expire in next 60 seconds, and if so, combines that timeout with the pseudotime synchronization timeout. That is, the rekey acts as both a TEK or KEK rekey and a pseudotime synchronization timeout rekey. See the "Time-Based Antireplay" section for more information on pseudotime synchronization.
Retransmitting a Rekey
Multicast rekeys are retransmitted by default. For unicast rekeys, if the key server does not receive the ACK, it retransmits the rekey. In either case, before retransmitting a rekey, the key server checks if there is a TEK or KEK rekey scheduled in the next 120 seconds. If so, it stops the current retransmission and waits for the scheduled rekey to happen.
Group Member Access Control List
For GET VPN, the traffic that has to be protected is defined statically on the key server using the ACL. The group member gets information about what has to be protected from the key server. This structure allows the key server to choose and change the policy dynamically as needed. In Secure Multicast, the key server ACL is defined inclusively. The ACL includes only the exact traffic that should be encrypted, with an implicit deny causing all other traffic to be allowed in the clear (that is, if there is no permit, all other traffic is allowed).
GET VPN employs a different philosophy: The definition of which packets should be encrypted is delivered independently. GET VPN supports only statically defined traffic selectors. Policy can be defined by using both deny and permit ACLs on the key server. Only the deny ACL is allowed to be manually configured on a group member. The policies that are downloaded from the key server and configured on the group member are merged. Any ACL that is configured on the group member has predominance over what is downloaded from the key server.
After the group member gets the ACL from the key server, the group member creates a temporary ACL and inserts it into the database. This ACL will be deleted if the group member is removed from the GDOI group for any reason. The packets that are going out of the interface are dropped by the group member if a packet matches the ACL but no IPsec SA exists for that packet.
The key server can send a set of traffic selectors, which may not exactly match the group member ACL on the group member. If such differences occur, the differences have to be merged and resolved. Because the group member is more aware of its topology than the key server, the downloaded ACLs are appended to the group member ACL. The group member ACL (except the implicit deny) is inserted into the database first, followed by the downloaded key server ACL. The database is prioritized, and the database search stops whenever a matched entry is found.
For information about configuring a group member ACL, see the "Configuring Group Member ACLs" section.
Behavior of a Group Member When Security Policy Changes
The behavior of a group member changes when ACL changes or any other policy changes are made in the key server. The effect of different policy changes on the behavior of the group members is explained in the following three scenarios.
Scenario 1
In the following example, the ACL has been initially configured to permit host A and host B.
ip access-list extended get-acl
permit ip host A host B
permit ip host B host A
Then the ACL is changed to permit host C and host D in the key server:
ip access-list extended get-acl
permit ip host C host D
permit ip host D host C
ACL changes affect the behavior of the group member in the following ways:
•Key server sends out a rekey to all group members immediately.
•Group member sends traffic between host A and host B in clear text immediately after rekey.
•Group member sends traffic between host C and host D in encrypted text immediately after rekey.
Scenario 2
The behavior of a group member changes when policy updates and transform set and TBAR changes are made to the key server.
In this scenario, it is assumed that:
•The transform set has been changed from ESP-3DES to ESP-AES.
•The policy change occurs at 1000 seconds before the current TEK lifetime expires.
These policy changes affect the behavior of the group member in the following ways:
•The key server sends out a rekey of both old SAs (3DES) and new SAs (AES).
•Group member continues to use the old SA (3DES) for 1000 seconds until it expires.
•After the old SA expires, the group member automatically switches over to new SAs (AES).
Scenario 3
The behavior of a group member changes when other policy updates in the key server involve both ACL changes and other changes like the transform set or TBAR.
In this scenario it is assumed that:
•The ACL has been updated as specified in Scenario 1.
•The transform set was changed from ESP-3DES to ESP-AES.
•The policy change occurs 1000 seconds before the current TEK lifetime expires.
ACL changes and other policy updates affect the behavior of the group member in the following ways:
•The key server sends out a rekey that consists of both old SAs (3DES) and new SAs (AES).
•The group member sends traffic between host A and host B in clear text immediately after rekey.
•The group member sends encrypted traffic between host C and host D using old SAs (3DES) for 1000 seconds until its TEK lifetime expires.
•When old SAs (3DES) expire, the group member automatically switches to new SAs to encrypt traffic between host C and host D in AES.
Fail-Close Mode
Until a group member registers with a key server, traffic passing through the group member is not encrypted. This state is called "fail open." To prevent unencrypted traffic from passing through a group member before that member is registered, you can configure the Fail-Close feature. If the feature is configured, an implicit "permit ip any any" policy is installed, and all unencrypted traffic passing through the group member is dropped (this state is called fail-close mode).
The fail-close function can also be achieved by configuring an interface ACL. However, the Fail-Close feature is more manageable and is easier to implement than ACL lists.
If you configure the Fail-Close feature, you can still allow specific unencrypted traffic to pass through the group member by configuring the match address command (match address {access-list-number | access-list-name}). This explicit "deny" ACL is added before the implicit "permit ip any any" to allow denied (unencrypted) traffic to pass through the group member.
After the group member has successfully completed its registration, the fail-close policy, both explicit and implicit, is removed, and the group member behaves as it did before the Fail-Close feature was configured.
Guidelines for Using the Fail-Close Feature
When you are configuring a crypto map to work in fail-close mode, you must be careful. If the fail-close ACL is defined improperly, you may lock yourself out of the router. For example, if you use Secure Shell (SSH) to log in to the router through the interface with the crypto map applied, you have to include the deny tcp any eq port host address command line under the fail-close ACL. You may also need to include the routing protocol that the router is using (such as deny ospf any any) to find the path to the key server. It is suggested that you configure fail-close and its ACL first, and then verify the fail-close ACL using the show crypto map gdoi fail-close map-name command. After you have checked your fail-close ACL and are confident that it is correct, you can make the crypto map work in the fail-close mode by configuring the activate command. Fail-close is not activated until you have configured the activate command.
The fail-close ACL is configured from the group-member perspective. The fail-close ACL is configured on group member as follows:
access-list 125 deny ip host host1-ip-addr host2-ip-addr
In fail-close mode, all IP traffic from host1 to host2 will be sent out by group member 1 in clear text. In addition, the inbound mirrored traffic (that is, IP traffic from host2 to host1) is also accepted by GM1 in clear text.
Note All IP traffic matching deny entries are sent out by the group member in clear text.
The inbound traffic is matched to the mirrored access list.
The fail-close access list follows the same rules as the group member access list. For more information, see the "Group Member Access Control List" section.
You need not configure the deny udp any eq 848 any eq 848 command to make the GDOI registration go through. The code itself checks whether it is a GDOI packet for a particular group member from the key server to which it is configured. If it is a GDOI packet for this group member, the packet is processed. But for a scenario in which the key server is behind group member 1, if group member 1 cannot register successfully with the key server, other group members also will not be able to register unless an explicit deny udp any eq 848 any eq 848 command line is configured for group member 1. However, if the Fail-Close feature is properly configured, even if a group member fails to register with a key server, you will be able to ensure that no unwanted traffic can go out "in the clear." But you can allow specified traffic to go out in the clear, in which case registration packets from other group members will be able to reach the key server through group member 1 even if it fails to get registered.
For information on configuring fail-close mode, see the "Activating Fail-Close Mode" section.
To verify whether fail-close mode is activated, use the show crypto map gdoi fail-close command.
Time-Based Antireplay
Antireplay is an important feature in a data encryption protocol such as IPSec (RFC 2401). Antireplay prevents a third party from eavesdropping on an IPsec conversation, stealing packets, and injecting those packets into a session at a later time. The time-based antireplay mechanism helps ensure that invalid packets are discarded by detecting the replayed packets that have already arrived at an earlier time.
GET VPN uses the Synchronous Antireplay (SAR) mechanism to provide antireplay protection for multisender traffic. SAR is independent of real-world Network Time Protocol (NTP) clock or sequential-counter mechanisms (which guarantee packets are received and processed in order). A SAR clock advances regularly. The time tracked by this clock is called pseudotime. The pseudotime is maintained on the key server and is sent periodically to the group members within a rekey message as a time-stamp field called pseudoTimeStamp. GET VPN uses a Cisco proprietary protocol called Metadata to encapsulate the pseudoTimeStamp. Group members have to be resynchronized to the pseudotime of the key server periodically. The pseudotime of the key server starts ticking from when the first group member registers. Initially, the key server sends the current pseudotime value of the key server and window size to group members during the registration process. New attributes, such as time-based replay-enabled information, window size, and the pseudotime of the key server, is sent under the SA payload (TEK).
The group members use the pseudotime to prevent replay as follows: the pseudoTimeStamp contains the pseudotime value at which a sender created a packet. A receiver compares the pseudotime value of senders with its own pseudotime value to determine whether a packet is a replayed packet. The receiver uses a time-based antireplay "window" to accept packets that contain a time-stamp value within that window. The window size is configured on the key server and is sent to all group members.
Note You should not configure time-based antireplay if you are using a Cisco VSA as a group member.
Figure 9 illustrates an antireplay window in which the value PTr denotes the local pseudotime of the receiver, and W is the window size.
Figure 9 Antireplay Window
Clock Synchronization
Clocks of the group members can slip and lose synchronization with the key server. To keep the clocks synchronized, a rekey message (multicast or unicast, as appropriate), including the current pseudotime value of the key server, is sent periodically, either in a rekey message or at a minimum of every 30 minutes to the group member. If a packet fails this antireplay check, the pseudotime of both the sender and receiver is printed, an error message is generated, and a count is increased.
To display antireplay statistics, use the show crypto gdoi group group-name gm replay command on both the sender and receiver devices. If the configuration is changed by the administrator to affect the replay method of the size configuration, the key server initiates a rekey message.
Interval Duration
A tick is the interval duration of the SAR clock. Packets sent in this duration have the same pseudoTimeStamp. The tick is also downloaded to group members, along with the pseudotime from the key server. For example, as shown in Figure 10, packets sent between T0 and T1 would have the same pseudoTimeStamp T0. SAR provides loose antireplay protection. The replayed packets are accepted if they are replayed during the window. The default window size is 100 seconds. It is recommended that you keep the window size small to minimize packet replay.
Figure 10 SAR Clock Interval Duration
Antireplay Configurations
The Antireplay feature can be enabled under IPsec SA on a key server by using the following commands:
•replay time window-size—Enables the replay time option, which supports the nonsequential, or time-based, mode. The window size is in seconds. Use this mode only if you have more then two group members in a group.
•replay counter window-size—Enables sequential mode. This mode is useful if only two group members are in a group.
•no replay counter window-size—Disables antireplay.
Control-Plane Time-Based Antireplay
Rekey Pseudotime Check
The rekey pseudotime check between key servers and group members is conducted as follows:
•The group member calculates the allowable pseudotime difference between the key server and its own as the lesser of the configured TBAR window size, that is, the value that was configured for it in the data plane, or 30 seconds.
–The group member accepts any rekey with a pseudotime larger than its own and updates its own pseudotime to the larger value. If the difference is larger than the calculated allowable pseudotime difference, it also generates the following syslog message:
*Jul 28 22:56:37.503: %GDOI-3-PSEUDO_TIME_LARGE: Pseudotime difference between key server (20008 sec) and GM (10057 sec) is larger than expected in group GET. Adjust to new pseudotime
–If the group member receives a rekey with a pseudotime smaller than its own but within the allowable difference, the group member accepts the rekey and updates its pseudotime value to the rekey pseudotime value.
–If the group member receives a rekey with a pseudotime smaller than its own but exceeding the allowable difference, the group member drops the rekey message and generates the following syslog message:
*Jul 28 23:37:59.699: %GDOI-3-PSEUDO_TIME_TOO_OLD: Rekey received in group GET is too old and fail PST check: my_pst is 22490 sec, peer_pst is 10026 sec, allowable_skew is 30 sec
ANN Message Pseudotime Handling in the Secondary Key Server
Cooperative key server announcement (ANN) messages are used to synchronize policy and group-member information between cooperative key servers.
The secondary key server handles ANN messages as follows:.
•The secondary key server calculates the allowable ANN message pseudotime as the lesser of the configured TBAR window size, that is, the value that was configured for it in the data plane, or 30 seconds.
•If the secondary key server receives an ANN message from the primary key server with a larger pseudotime, it does the following:
1. It updates its pseudotime to the primary key server's value.
2. If the pseudotime difference is larger than allowable, it generates the following syslog message:
*Jul 28 23:48:56.871: %GDOI-4-GDOI_ANN_TIMESTAMP_LARGE: COOP_KS ANN received from KS 10.0.8.1 in group GET has pseudotime bigger than myself. Adjust to new pseudotime:
my_old_pst is 23147 sec, peer_pst is 30005 sec
•If the secondary key server receives an ANN message from the primary key server with a smaller pseudotime, it behaves as follows:
–If the difference is within the allowable range, the secondary key server accepts it and updates its pseudotime to the primary key server's value.
–If the difference exceeds the allowable range, it generates the following syslog message:
*Jul 28 23:42:12.603: %GDOI-4-GDOI_ANN_TIMESTAMP_TOO_OLD: COOP_KS ANN from KS 10.0.8.1 in group GET is too old and fail PST check:
my_pst is 22743 sec, peer_pst is 103 sec, allowable_skew is 10 sec
If, after three retransmit requests, the secondary key server has still not received any ANN message with a valid pseudotime, it starts blocking new group-member registrations, as follows:
*Jul 28 23:38:57.859: %GDOI-5-COOP_KS_VALID_ANN_TIMER_EXPIRED: This sec-KS has NOT received an ANN with valid pseudotime for an extended period in group GET. It will block new group members registration temporarily until a valid ANN is received
*Jul 29 00:08:47.775: %GDOI-5-COOP_KS_BLOCK_NEW_GM_REGISTER: This key server temporarily blocks group member with ip-addr 10.0.0.2 from registering in group GET as it has not received an ANN with valid pseudotime for prolonged period
The secondary key server resumes its group member registration functionality if any of the following happens:
•It receives an ANN with a valid pseudotime from the primary key server.
•It becomes a primary key server itself.
•The clear crypto gdoi group command is executed on the secondary key server.
ANN Message Pseudotime Handling in the Primary Key Server
The primary key server handles ANN messages as follows:
•It calculates the allowable ANN message pseudotime as the lesser of the configured TBAR window size, that is, the value that was configured for it in the data plane, or 30 seconds.
•It accepts from the secondary key server ANN messages that have a smaller pseudotime but are within the allowable difference.
•It rejects ANN messages that have a smaller pseudotime but exceed the allowable difference.
During a network merge, the following conditions apply:
•The new primary key server always picks the larger pseudotime between the two key servers.
•If the difference is larger than the calculated allowable pseudotime difference, the new primary key server sends out rekeys to all group members to update their pseudotime. It also generates the following syslog messages:
*Jul 28 23:42:41.311: %GDOI-5-COOP_KS_ELECTION: KS entering election mode in group GET (Previous Primary = NONE)
*Jul 28 23:42:41.311: %GDOI-4-GDOI_ANN_TIMESTAMP_LARGE: COOP_KS ANN received from KS 10.0.9.1 in group GET has PST bigger than myself. Adjust to new pseudotime:
my_old_pst is 0 sec, peer_pst is 22772 sec
*Jul 28 23:43:16.335: %GDOI-5-COOP_KS_TRANS_TO_PRI: KS 10.0.8.1 in group GET transitioned to Primary (Previous Primary = NONE)
*Jul 28 23:43:16.347: %GDOI-5-KS_SEND_UNICAST_REKEY: Sending Unicast Rekey for group GET from address 10.0.8.1 with seq # 1
Cooperative Key Server
Figure 11 illustrates cooperative key server key distribution. The text following the illustration explains the Cooperative Key Server feature.
Figure 11 Cooperative Key Server Key Distribution
Cooperative key servers provide redundancy to GET VPN. Multiple key servers are supported by GET VPN to ensure redundancy, high availability, and fast recovery if the primary key server fails. Cooperating GDOI key servers jointly manage the GDOI registrations for the group. Each key server is an active key server, handling GDOI registration requests from group members. Because the key servers are cooperating, each key server distributes the same state to the group members that register with it. Load balancing is achieved because each of the GDOI key servers can service a portion of the GDOI registrations.
The primary key server is responsible for creating and distributing group policy. When cooperative key server key distribution occurs, one key server declares itself as primary, creates a policy, and sends the policy to the other secondary key server. The secondary key server declares the primary key server as primary key server when it gets the policy and ends the election mode. The secondary key server now also blocks GM registration while the cooperative key server key distribution is in progress. This change allows the cooperative key server distribution to become more efficient because it saves time. For example, the syslog warning message similar to the following is displayed during distribution:
00:00:16: %GDOI-5-COOP_KS_BLOCK_NEW_GM_REGISTER_ELECTION: This KS temporarily blocks GM with ip-addr 10.0.4.1 from registering in group diffint as the KS election is underway
The primary key server periodically sends out (or broadcasts) group information updates to all other key servers to keep those servers in synchronization. If the secondary key servers somehow miss the updates, they contact the primary key server to directly request information updates. The secondary key servers mark the primary key server as unreachable (that is, "dead") if the updates are not received for an extended period.
When a new policy is created on a primary key server, regardless of which key server a group member may be registered with, it is the responsibility of the primary key server to distribute rekey messages to GDOI group members.
In a cooperative-key-server setting, the rekey sequence number is synchronized between the primary and secondary key servers.
In a network merge, the key servers pick the larger of the rekey sequence numbers that they have between them.
If you are supporting more than 300 group members in your cooperative key server setup, you should increase the buffer size by using the buffers huge size command.
Announcement Messages
Announcement messages are secured by IKE Phase 1 and are sent as IKE notify messages. Authentication and confidentiality that are provided by IKE is used to secure the messaging between the key servers. Antireplay protection is provided by the sequence numbers in the announcement messages. Announcement messages are periodically sent from primary to secondary key servers.
Announcement messages include the components, described in the following sections that help maintain the current state.
Sender Priority of a Key Server
This value describes the priority of the sender, which is configurable using the CLI. The key server with the highest priority becomes the primary key server. If the priority values are the same, the key server with the highest IP address becomes the primary key server.
Maintaining the Role of the Sender
During the synchronization period, if the key servers are at geographically dispersed locations, they may suffer a network-partitioning event. If a network-partitioning event occurs, more than one key server can become the primary key server for a period of time. When the network is operating normally again and all the key servers find each other, they need to be told the current role of the sender so the key servers can attain their proper roles.
Request for a Return Packet Flag
All messages are defined as one-way messages. When needed, a key server can request the current state from a peer to find out its role or request the current state of the group.
Group Policies
The group policies are the policies that are maintained for a group, such as group member information and IPsec SAs and keys.
Antireplay functionalities and incorporated Cooperative announcement messages are supported. The primary key server updates the pseudotime value, sending it to all secondary key servers in the group. The secondary key servers should synchronize their SAR clocks to this updated value.
ANN Message Sequence Number Check Between Cooperative Key Servers
The following describes the sequence number check between cooperative key servers:
•Cooperative key servers drop any ANN message with a sequence number smaller than or equal to that of the last received ANN message.
•The ANN message is accepted if the sequence number is larger than that of the last received rekey message, no matter how large the difference.
•If a key server is reloaded, a new IKE session is created between the peers, and the reloaded key server's ANN sequence number will start with zero. In this case, the other side will accept the ANN message with any sequence number.
Change Key Server Role
In a network of cooperative key servers, the primary server is elected based on its highest priority at the time of election. The other key servers have secondary status. If the primary key server is detected as being dead or if its role changes, the clear crypto gdoi ks cooperative role command allows you to reset the cooperative role of the primary key server.
If the clear crypto gdoi ks cooperative role command is executed on a secondary key server, the election is triggered on that secondary key server although that server would most likely remain a secondary key server because there has been an elected primary key server. However, if the clear crypto gdoi ks cooperative role command is executed on the primary key server, the primary key server is reassigned to a secondary role, and as a result, a new election that involves all the key servers is triggered. If the previous primary server has the highest priority (of all the key servers), it again becomes the primary server. If the previous primary server does not have the highest priority, the server having the highest priority is elected as the new primary server.
Receive Only SA
For multicast traffic using the GDOI protocol, bidirectional SAs are installed. The Receive Only feature enables an incremental deployment so that only a few sites can be verified before bringing up an entire network. To test the sites, one of the group members should send encrypted traffic to all the other group members and have them decrypt the traffic and forward the traffic "in the clear." Receive Only SA mode allows encryption in only the inbound direction for a period of time. (See the steps for the Receive Only SA process.) If you configure the sa receive-only command on the key server, Steps 2 and 3 happen automatically.
1. Mark IPsec SAs as "receive-only" on the GDOI key server.
This action allows the group members to install SAs in the inbound direction only. Receive-only SAs can be configured under a crypto group. (See the "Configuring the Group ID, Server Type, and SA Type" section.)
2. Mark GDOI TEK payloads as "receive only."
If the sa receive-only command is configured, all TEKs under this group are going to be marked "receive only" by the key server when they are sent to the group member.
3. Install one-way IPsec flows.
Every time a GDOI group member receives an IPsec SA from the key server that is marked as "receive only," the group member installs this IPsec SA only in the inbound direction rather than in both incoming and outgoing directions.
4. Test individual group members using the following local-conversion commands:
•crypto gdoi gm ipsec direction inbound optional
•crypto gdoi gm ipsec direction both
First, individually convert each of the group members to passive mode (this change tells the outbound check that there is a valid SA) and then to bidirectional mode.
5. Globally convert from "receive only" to "receive and send."
The following method can be used when the testing phase is over and "receive only" SAs have to be converted to bidirectional SAs.
Global Conversion
Remove the sa receive-only command under the group. Removing the sa receive-only command creates new IPsec SAs for this group and causes a rekey. On receipt, group members reinstall the SA in both directions and begin to use it in passive mode. Because the SA cannot remain in passive mode forever, the group members change those SAs to receive or send mode if there is no rekey in 5 minutes. The conversion from passive mode to bidirectional encryption mode is automatic and does not require the administrator to do anything.
Passive SA
The Passive SA feature allows you to configure a group member so that it is in passive mode permanently. By using the Passive SA feature, you will avoid having to use the crypto gdoi gm ipsec direction inbound optional privileged EXEC command, which is not persistent after a router reload and can be overridden by key server configuration from a rekey. Having the group member in passive mode benefits network testing and debugging during migration to GET VPN, and it provides complete encryption protection during the migration. The group-member passive-mode configuration has higher priority over a key server configuration. The crypto gdoi gm ipsec direction inbound optional privileged EXEC command can override the configuration until the next rekey, which will bring back the group member and key server configuration.
To configure the Passive SA feature, see the "Configuring Passive SA" section.
Enhanced Solutions Manageability
Several show and debug commands are supported to help verify functionality. See the "Activating Fail-Close Mode" section for details.
Support with VRF-Lite Interfaces
The VRF-Lite application supports segmentation of traffic in the control and forwarding planes by keeping the routing tables separate for each user group (or VPN) and forwarding the traffic on the associated or dedicated interfaces of each user group.
There are some deployment scenarios in which remote sites that are connecting to an MPLS VPN network might be extending segmentation from a campus to the WAN. In such an extended segmentation case, a CE-PE interface on a CE (group member or key server) device "bounds" to its associated virtual routing and forwarding (VRF) instance. This VRF interface connects to an MPLS PE device where it is directly mapped to its associated Border Gateway Protocol (BGP) VRF process, in which case the crypto map is applied to a VRF interface. No other configuration changes are necessary.
GET VPN VRF-Aware GDOI on GM
The GET VPN VRF-Aware GDOI on GM feature adds to the GET VPN feature. The GET VPN VRF-Aware GDOI on GM feature allows control plane traffic (for example, GDOI registration) to be placed in a separate VRF (for example, a dedicated management VRF). Multiple entities placed in different MPLS VPNs (VRFs) can share the same key servers.
Use the client registration interface command to allow GDOI crypto maps to route registrations and receive rekeys through a different VRF.
The GET VPN GM requires VRF-lite in its data path. VRF-lite means that all the traffic traversing in a particular VRF should use the same VRF after route lookup. VRF-lite does not cover route leaking. If route leaking is configured on the GM, packets will be sent out in clear text from the crypto map applied interface.
The scenarios are described in the following sections.
Shared GDOI Groups/Policies and Crypto Maps
The same crypto map is applied to multiple subinterfaces, and each subinterface is in a different VRF context or the same VRF context, and the key server is accessible through a different VRF or global VRF table. This is addressed by separating out the control traffic.
There is one group member registration for all the crypto maps applied to different subinterfaces. After successful registration, policies are downloaded to where the crypto maps are.
Different Groups/Policies with Different Crypto Maps Sharing a Key Server
Different groups are applied to different interfaces, and each of these interfaces is in a different VRF context or the same VRF context. All these groups are accessing the same key server, and this key server is accessible through a different VRF or global VRF table. This is addressed by having individual group members registering for each group.
So for every group, there is one registration and also one unicast rekey received. All these group members would be using the same IKE SA, which is established between the CE and the key server.
Different Groups/Policies with Different Crypto Maps on Different Key Servers
Different groups are applied to different interfaces, and each of these interfaces is in a different VRF context or the same VRF context. All these groups are accessing different key servers, and these key servers are accessible through a different VRF or the global VRF table. This is addressed by having a group member registering for each group. So for every group, there is one registration and also one unicase rekey received. Because every registration is with a different key server, a different IKE SA is established for every registration.
Authentication Policy for GM Registration
GMs can authenticate to the key server at registration time using preshared keys or Public Key Infrastructure (PKI). Preshared keys are easy to deploy but must be managed proactively. We recommend that you deploy a peer-based preshared key instead of defining a default key (the key defined with an address of 0.0.0.0) for all the devices in the network. Preshared keys should be updated regularly (every few months).
Note A preshared key can be updated on a key server-group member (KS-GM) peer basis without affecting the crypto data plane or control plane because rekeys are secured using the KEK. It is important to ensure that a GM can reregister to each ordered set of key servers using the newly assigned preshared key.
PKI uses its infrastructure to overcome the key management difficulties encountered when preshared keys are used. The PKI infrastructure acts as a certificate authority (CA) where router certificates are issued and maintained. However, using PKI during IKE authentication is computationally intensive. In PKI deployments, key server capacity, design, and placement become important.
For added security, GET VPN also supports GM authorization using either preshared keys or PKI. For more information, see the "GET VPN GM Authorization" section.
GET VPN GM Authorization
GET VPN GM authorization can be done using preshared keys or PKI. It is a best practice to turn on GET VPN authorization. When a key server serves multiple GDOI groups, key server authorization is required to prevent a GM in one group from requesting keys and policies from another group. The ISAKMP authorization confirms that the GM is allowed to request GDOI attributes from the key server while the GDOI authorization confirms that the GM is allowed to request GDOI attributes from a specific group configured in the key server.
GM authorization using preshared keys or PKI is discussed in the following sections:
•GM Authorization Using Preshared keys
GM Authorization Using Preshared keys
GET VPN supports GM authorization using the IP address when preshared keys are used. An ACL matching the WAN addresses (or subnets) of the GM can be defined and applied to the GET VPN group configuration. Any GM whose IP addresses match the ACL authorizes successfully and can register to the key server. If a GM IP address does not match the ACL, the key server rejects the GM registration request.
In cases of unsuccessful authorization, the following syslog message is generated:
%GDOI-1-UNAUTHORIZED_IPADDR: Group getvpn received registration from
unauthorized ip address: 10.1.1.9
GM Authorization Using PKI
GET VPN supports GM authorization using the commonly used distinguished name (DN) or fully qualified domain name (FQDN) when PKI is used. The authorization identity command is used to activate GN authorization. A crypto identity matching certain fields in the GM certificate (typically—organizational unit (OU)) can be defined and applied to the GET VPN group configuration. Use the crypto identity command to define a crypto identity.
Any GM whose certificate credentials match the ISAKMP identity is authorized and can register to the key server. For example, if all GM certificates are issued with OU=GETVPN, a key server can be configured to check (authorize) that all GMs present a certificate having OU=GETVPN. If the OU in the certificate presented by a GM is set to something else, the GM will not be authorized to register to the key server.
If authorization is unsuccessful, the following syslog message is generated:
%GDOI-1-UNAUTHORIZED_IDENTITY: Group getvpn received registration from
unauthorized identity: Dist.name: hostname=GroupMember-1, ou=TEST
Rekey Functionality in Protocol Independent Multicast-Sparse Mode
Multicast rekeying can be used with all modes of multicast. The rekey retransmit command should be used whenever the Protocol Independent Multicast-sparse mode (PIM-SM) is configured because the PIM-SM shortest path tree (SPT) can be torn down if it does not receive continuing traffic. When traffic resumes, PIM-SM must reestablish the SPT. Retransmitting rekey packets increases the chance that group members receive the rekeys when PIM-SM is setting up the SPT.
How to Configure Cisco Group Encrypted Transport VPN
This section includes the following required and optional tasks:
•Configuring a Key Server (required)
•Configuring a Group Member (required)
•Configuring GET VPN GM Authorization (optional)
•Verifying and Troubleshooting Cisco Group Encrypted Transport VPN Configurations (optional)
Configuring a Key Server
To configure a key server, perform the steps in the following subtasks.
•Configuring RSA Keys to Sign Rekey Messages (optional)
•Configuring the Group ID, Server Type, and SA Type (required)
•Configuring the Rekey (optional)
•Configuring Group Member ACLs (optional)
•Configuring an IPsec Lifetime Timer (optional)
•Configuring an ISAKMP Lifetime Timer (optional)
•Configuring the IPsec SA (required)
•Configuring Time-Based Antireplay for a GDOI Group (optional)
•Configuring Passive SA (optional)
•Resetting the Role of the Key Server (optional)
Prerequisites
Before creating the GDOI group, you must first configure IKE and the IPsec transform set, and you must create an IPsec profile. For information about how to configure IKE and the IPsec transform set and to create an IPsec profile, see the "Related Documents" subsection of the "Additional References" section.
Configuring RSA Keys to Sign Rekey Messages
To configure RSA keys that will be used to sign rekey messages, perform the following steps. Omit this subtask if rekey is not in use.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto key generate rsa general-keys label name-of-key
DETAILED STEPS
What to Do Next
Configure the group ID, server type, and SA type. (See the "Configuring the Group ID, Server Type, and SA Type" section.)
Configuring the Group ID, Server Type, and SA Type
For a large number of sites, it is better to take precautions and add functionality incrementally, especially when migrating from any other encryption solutions like Dual Multipoint VPN (DMVPN). For example, instead of setting up all the CPE devices to encrypt the traffic bidirectionally, it is possible to configure one-way encryption so that only one or fewer members of a group are allowed to send encrypted traffic. Others are allowed to receive only encrypted traffic. After the one-way encryption is validated for one or a few members, bidirectional encryption can be turned on for all the members. This "inbound only" traffic can be controlled using the sa receive only command under a crypto group.
To configure the group ID, server type, and SA type, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto gdoi group group-name
4. identity number number
or
identity address ipv4 address
5. server local
6. sa receive-only
DETAILED STEPS
What to Do Next
Remove the receive-only configuration on the key server so that the group members are now operating in bidirectional receive and send mode.
Configuring the Rekey
This section includes the following optional tasks:
•Configuring a Multicast Rekey
Rekey is used in the control plane by the key server to periodically refresh the policy and IPsec SAs of the group. On the group-member side, instead of fully reregistering when timers expire for any other reasons, refreshing the registration with a rekey is more efficient. The initial registration is always a unicast registration.
The key server can be configured to send rekeys in unicast or multicast mode. The rekey transport mode is determined by whether the key server can use IP multicast to distribute the rekeys. If multicast capability is not present within the network of the customer, the key server will have to be configured to send rekeys using unicast messages.
Additional options for rekey use the rekey authentication, rekey retransmit, and rekey address ipv4 commands. If unicast transport mode is configured, the source address command will have to be included to specify the source address of this unicast rekey message.
Multicast is the default transport type for rekey messages. The following bulleted items explain when to use rekey transport type multicast or unicast:
•If all members in a group are multicast capable, do not configure the rekey transport unicast command. The no rekey transport unicast command is not needed if the rekey transport type "unicast" was not configured previously under this group because multicast rekeys are on by default.
•If all members in a group are unicast, use the rekey transport unicast command.
•If you have mixed members in a group (that is, the majority are multicast, but a few are unicast), do not configure the rekey transport unicast command. The rekeys will be distributed using multicast to the majority of group members. The remainder of the group members that do not receive the multicast messages (unicast group members) will have to reregister to the key server when their policies expire. Mixed mode (that is, unicast and multicast rekey mode) is not supported.
If the no rekey transport unicast command is used, members in the GDOI group that are unable to receive the multicast rekey messages need to reregister with the key server to get the latest group policies. The reregistering forces the default transport type to multicast. If no transport type was configured previously, the multicast transport type will apply by default.
Prerequisites
Before configuring the rekey authentication command, you must have configured the router to have an RSA key generated using the crypto key generate rsa command and general-keys and label keywords (for example, "crypto key generate rsa general-key label my keys").
Configuring a Unicast Rekey
In the configuration task table, the address "ipv4 10.0.5.2" specifies the interface on the key server by which the unicast or multicast rekey messages are sent. This address is required for unicast rekeys, but it is optional for multicast rekeys. For multicast rekeys, the source address of the key server can be retrieved from the rekey ACL.
To configure a unicast rekey, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto gdoi group group-name
4. identity number number
or
identity address ipv4 address
5. server local
6. rekey transport unicast
7. rekey lifetime seconds number-of-seconds
8. rekey retransmit number-of-seconds number number-of-retransmissions
9. rekey authentication {mypubkey | pubkey} rsa key-name
10. address ipv4 ipv4-address
DETAILED STEPS
Configuring a Multicast Rekey
To configure a multicast rekey, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto gdoi group group-name
4. identity number number
or
identity address ipv4 address
5. server local
6. rekey address ipv4 {access-list-name | access-list-number}
7. rekey lifetime seconds number-of-seconds
8. rekey retransmit number-of-seconds number number-of-retransmissions
9. rekey authentication {mypubkey | pubkey} rsa key-name
10. exit
11. exit
12. access-list access-list-number {deny | permit} udp host source [operator [port]] host source [operator [port]]
DETAILED STEPS
Configuring Group Member ACLs
All IP traffic matching deny entries are sent out by the group member in clear text. The inbound traffic is matched to the mirrored access list.
To configure group member ACLs, perform this task (note that a group member access list can contain only deny statements).
SUMMARY STEPS
1. enable
2. configure terminal
3. access-list access-list-number deny ip host source host source
4. access-list access-list-number permit ip source
DETAILED STEPS
What to Do Next
The access list defined in Step 4 is the same one that should be used to configure the SA. See the "Configuring the IPsec SA" section.
Configuring an IPsec Lifetime Timer
To configure an IPsec lifetime timer for a profile, perform the following steps. If this configuration task is not performed, the default is the maximum IPsec SA lifetime of 3600 seconds. The TEK lifetime value should be more than 900 seconds.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto ipsec profile name
4. set security-association lifetime seconds seconds
DETAILED STEPS
What to Do Next
Configure the IPsec SA. See the "Configuring the IPsec SA" section.
Configuring an ISAKMP Lifetime Timer
To configure an ISAKMP lifetime timer, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto isakmp policy priority
4. lifetime seconds
DETAILED STEPS
Configuring the IPsec SA
If time-based antireplay is configured on the key server but the group member is not capable of supporting it, the GDOI-3-GM_NO_CRYPTO_ENGINE syslog message is logged to the group member. See Appendix I: System Logging Messages for a list of system error messages.
To configure the IPsec SA, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto gdoi group group-name
4. identity number number
or
identity address ipv4 address
5. server local
6. sa ipsec sequence-number
7. profile ipsec-profile-name
8. match address ipv4 {access-list-number | access-list-name}
9. exit
10. exit
11. exit
12. crypto ipsec transform-set transform-set-name transform [transform2...transform4]
13. crypto ipsec profile ipsec-profile-name
14. set transform-set transform-set-name
DETAILED STEPS
What to Do Next
Replay should be configured. If replay is not configured, the default is counter mode.
Configuring Time-Based Antireplay for a GDOI Group
To configure time-based antireplay for a GDOI group, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto gdoi group group-name
4. identity number policy-name
5. server local
6. address ip-address
7. sa ipsec sequence-number
8. profile ipsec-profile-name
9. match address {ipv4 access-list-number | access-list-name}
10. replay counter window-size seconds
11. replay time window-size seconds
DETAILED STEPS
Configuring Passive SA
To configure passive SA (to put the group member in passive mode), perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto gdoi group group-name
4. identity name
5. passive
6. server address ipv4 {address | hostname}
DETAILED STEPS
Resetting the Role of the Key Server
To reset the cooperative role of the primary key server, perform the following steps.
SUMMARY STEPS
1. enable
2. clear crypto gdoi ks cooperative role
DETAILED STEPS
Configuring a Group Member
To configure a group member, perform the following subtasks:
•Configuring the Group Name, ID, Key Server IP Address, and Group Member Registration (required)
•Creating a Crypto Map Entry (required)
•Applying the Crypto Map to an Interface to Which the Traffic Must Be Encrypted (required)
•Activating Fail-Close Mode (optional)
•Configuring Ciphers or Hash Algorithms for KEK (optional)
•Configuring Acceptable Transform Sets for TEK (optional)
Configuring the Group Name, ID, Key Server IP Address, and Group Member Registration
To configure the group name, ID, key server IP address, and group member registration, perform the following steps. You can configure up to eight key server addresses.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto gdoi group group-name
4. identity number number
or
identity address ipv4 address
5. server address ipv4 address
6. client registration interface
DETAILED STEPS
What to Do Next
Configure a crypto map. See the "Creating a Crypto Map Entry" section.
Creating a Crypto Map Entry
To create a crypto map entry and associate a GDOI group to it, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto map map-name seq-num gdoi
4. set group group-name
DETAILED STEPS
What to Do Next
Apply the crypto map to an interface to which the traffic has to be encrypted. See the "Applying the Crypto Map to an Interface to Which the Traffic Must Be Encrypted" section.
Applying the Crypto Map to an Interface to Which the Traffic Must Be Encrypted
To apply the crypto map to an interface to which the traffic must be encrypted, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. interface type slot/port
4. crypto map map-name redundancy standby-group-name stateful
DETAILED STEPS
Activating Fail-Close Mode
To configure a crypto map to work in fail-close mode, perform the following steps:
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto map map-name gdoi fail-close
4. match address access-list-number
5. activate
DETAILED STEPS
Configuring Ciphers or Hash Algorithms for KEK
To configure acceptable ciphers or hash algorithms for KEK, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto gdoi group group-name
4. identity number number
or
identity address ipv4 address
5. server address ipv4 address
6. client rekey encryption cipher [...[cipher]]
7. client rekey hash hash
8. end
DETAILED STEPS
Configuring Acceptable Transform Sets for TEK
To configure the acceptable transform sets used by TEK for data encryption or authentication, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto ipsec transform-set transform-set-name transform [transform2...transform4]
4. exit
5. crypto gdoi group group-name
6. client transform-sets transform-set-name1 [...[transform-set-name6]]
7. end
DETAILED STEPS
Configuring GET VPN GM Authorization
GET VPN GM authorization can be done using preshared keys or PKI. It is a best practice to turn on GET VPN authorization. When a key server serves multiple GDOI groups, key server authorization is required to prevent a GM in one group from requesting keys and policies from another group. The ISAKMP authorization confirms that the GM is allowed to request GDOI attributes from the key server while the GDOI authorization confirms the GM is allowed to request GDOI attributes from a specific group configured in the key server.
To configure GET VPN GM authorization, perform either of the following tasks:
•Configuring GM Authorization Using Preshared Keys (optional)
•Configuring GM Authorization Using PKI (optional)
Configuring GM Authorization Using Preshared Keys
To configure GM authorization using preshared keys, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto gdoi group group-name
4. server local
5. authorization address ipv4 {access-list-name | access-list-number}
6. exit
7. exit
8. access-list access-list-number [dynamic dynamic-name [timeout minutes]] {deny | permit} protocol source source-wildcard destination destination-wildcard [precedence precedence] [tos tos] [time-range time-range-name] [fragments] [log [word] | log-input [word]]
9. exit
DETAILED STEPS
Configuring GM Authorization Using PKI
To configure GM authorization using PKI, perform the following steps.
SUMMARY STEPS
1. enable
2. configure terminal
3. crypto isakmp identity {address | dn | hostname}
4. crypto pki trustpoint name
5. subject-name [x.500-name]
6. exit
7. crypto gdoi group group-name
8. server local
9. authorization identity name
10. exit
11. exit
12. crypto identity name
13. dn name=string [, name=string]
14. exit
15. crypto isakmp identity {address | dn | hostname}
16. crypto pki trustpoint name
17. subject-name [x.500-name]
18. end
DETAILED STEPS
Verifying and Troubleshooting Cisco Group Encrypted Transport VPN Configurations
The following tasks can be used to verify and troubleshoot your GET VPN configurations. These tasks are optional and are used to gather information during troubleshooting.
•Verifying Active Group Members on a Key Server
•Verifying Rekey-Related Statistics
•Verifying IPsec SAs That Were Created by GDOI on a Group Member
•Verifying Cooperative Key Server States and Statistics
•Verifying Antireplay Pseudotime-Related Statistics
•Verifying the Fail-Close Mode Status of a Crypto Map
Verifying Active Group Members on a Key Server
To verify active group members on a key server, perform the following steps.
SUMMARY STEPS
1. enable
2. show crypto gdoi ks members
DETAILED STEPS
Verifying Rekey-Related Statistics
To verify rekey-related statistics, perform the following steps.
SUMMARY STEPS
1. enable
2. show crypto gdoi ks rekey
DETAILED STEPS
Verifying IPsec SAs That Were Created by GDOI on a Group Member
To verify IPsec SAs that were created by GDOI on a group member, perform the following steps.
SUMMARY STEPS
1. enable
2. show crypto gdoi group group-name ipsec sa
DETAILED STEPS
Verifying IPsec SAs That Were Created by GDOI on a Key Server
To verify IPsec SAs that were created by GDOI on a key server, perform the following steps.
SUMMARY STEPS
1. enable
2. show crypto ipsec sa
DETAILED STEPS
Verifying Cooperative Key Server States and Statistics
To verify cooperative key server states and statistics, perform the following steps, using one or both of the debug and show commands shown.
SUMMARY STEPS
1. enable
2. debug crypto gdoi ks coop
3. show crypto gdoi group group-name ks coop [version]
DETAILED STEPS
Verifying Antireplay Pseudotime-Related Statistics
To verify antireplay pseudotime-related statistics, perform the following steps using one or all of the clear, debug, and show commands.
SUMMARY STEPS
1. enable
2. clear crypto gdoi group group-name replay
3. debug crypto gdoi replay
4. show crypto gdoi group group-name
5. show crypto gdoi group group-name ks replay
DETAILED STEPS
Verifying the Fail-Close Mode Status of a Crypto Map
To verify the fail-close mode status of a crypto map, perform the following steps.
SUMMARY STEPS
1. enable
2. show crypto map gdoi fail-close
DETAILED STEPS
Configuration Examples for Cisco Group Encrypted Transport VPN
This section includes the following case study and configuration examples:
•Example: Key Server and Group Member Case Study
Example: Key Server and Group Member Case Study
The following case study includes encrypting traffic CE-CE in an MPLS VPN environment.
The MPLS VPN core interconnects VPN sites as is shown in Figure 12. VPN site CPEs, Group Member 1 through Group Member 4, are grouped into a single GDOI group that correlates with a VPN with which these sites are a part. This scenario is an intranet VPN scenario. All the key servers and Group Members are part of the same VPN. Key Server 1 and Key Server 2 are the cooperative key servers that support VPN members Group Member 1 through Group Member 4. Key Server 1 is the primary key server and Key Server 2 is the secondary key server.
Figure 12 Key Server and Group Member Scenario
The following configuration examples are based on the case study in Figure 12.
Example: Key Server 1
Key server 1 is the primary key server.
service timestamps debug datetime msec
service timestamps log datetime msec
no service password-encryption
service internal
!
hostname KS1
!
logging buffered 100000 debugging
no logging console
!
no aaa new-model
!
resource policy
!
clock timezone EST 0
ip subnet-zero
no ip domain lookup
ip domain name cisco.com
!
crypto isakmp policy 1
encr 3des
authentication pre-share
group 2
lifetime 400
crypto isakmp key cisco address 10.1.1.13
crypto isakmp key cisco address 10.1.1.9
crypto isakmp key cisco address 10.1.1.1
crypto isakmp key cisco address 10.1.1.5
crypto isakmp key cisco address 10.1.1.21
!
crypto ipsec transform-set gdoi-trans-group1 esp-3des esp-sha-hmac
!
crypto ipsec profile gdoi-profile-group1
set security-association lifetime seconds 1800
set transform-set gdoi-trans-group1
!
crypto gdoi group group1
identity number 1
server local
rekey lifetime seconds 86400
rekey retransmit 10 number 2
rekey authentication mypubkey rsa group1-export-general
rekey transport unicast
sa ipsec 1
profile gdoi-profile-group1
match address ipv4 101
replay counter window-size 64
address ipv4 209.165.200.225
redundancy
local priority 10
peer address ipv4 209.165.200.225
!
interface Ethernet0/0
ip address 209.165.200.225 255.255.255.252
!
ip classless
ip route 0.0.0.0 0.0.0.0 10.1.1.18
!
access-list 101 permit ip 10.1.0.0 0.0.255.255 10.1.0.0 0.0.255.255
!
end
Example: Key Server 2
Key Server 2 is the secondary key server.
service timestamps debug datetime msec
service timestamps log datetime msec
no service password-encryption
service internal
!
hostname KS2
!
logging buffered 100000 debugging
no logging console
!
no aaa new-model
!
resource policy
!
clock timezone EST 0
ip subnet-zero
no ip domain lookup
ip domain name cisco
!
crypto isakmp policy 1
encr 3des
authentication pre-share
group 2
lifetime 400
crypto isakmp key cisco address 10.1.1.9
crypto isakmp key cisco address 10.1.1.1
crypto isakmp key cisco address 10.1.1.5
crypto isakmp key cisco address 10.1.1.17
crypto isakmp key cisco address 10.1.1.13
!
crypto ipsec transform-set gdoi-trans-group1 esp-3des esp-sha-hmac
!
crypto ipsec profile gdoi-profile-group1
set security-association lifetime seconds 1800
set transform-set gdoi-trans-group1
!
crypto gdoi group group1
identity number 1
server local
rekey lifetime seconds 86400
rekey retransmit 10 number 2
rekey authentication mypubkey rsa group1-export-general
rekey transport unicast
sa ipsec 1
profile gdoi-profile-group1
match address ipv4 101
replay counter window-size 64
address ipv4 10.1.1.21
redundancy
local priority 1
peer address ipv4 10.1.1.17
!
interface Ethernet0/0
ip address 209.165.200.225 255.255.255.252
!
ip classless
ip route 0.0.0.0 0.0.0.0 10.1.1.22
!
access-list 101 permit ip 10.1.0.0 0.0.255.255 10.1.0.0 0.0.255.255
!
end
Example: Group Member 1
Group member 1 is part of a GDOI group that correlates with a VPN with which these sites are a part.
service timestamps debug datetime msec
service timestamps log datetime msec
no service password-encryption
!
hostname GM1
!
resource policy
!
clock timezone EST 0
ip subnet-zero
!
crypto isakmp policy 1
encr 3des
authentication pre-share
group 2
lifetime 3600
crypto isakmp key cisco address 209.165.200.225
crypto isakmp key cisco address 209.165.201.1
!
crypto gdoi group group1
identity number 1
server address ipv4 209.165.200.225
server address ipv4 209.165.201.1
!
crypto map map-group1 10 gdoi
set group group1
!
interface Ethernet0/0
ip address 209.165.200.225 255.255.255.252
crypto map map-group1
!
router bgp 1000
no synchronization
bgp log-neighbor-changes
network 10.1.1.0 mask 255.255.255.0
neighbor 10.1.1.2 remote-as 5000
no auto-summary
!
ip classless
!
End
The same GDOI group cannot be applied to multiple interfaces. The following examples show unsupported cases:
Example 1
crypto map map-group1
group g1
interface ethernet 1/0
crypto map map-group1
interface ethernet 2/0
crypto map map-group1
Example 2
crypto map map-group1 10 gdoi
set group group1
crypto map map-group2 10 gdoi
set group group1
interface ethernet 1/0
crypto map map-group1
interface ethernet 2/0
Example: Group Member 2
Group member 2 is part of a GDOI group that correlates with a VPN with which these sites are a part.
service timestamps debug datetime msec
service timestamps log datetime msec
!
hostname GM2
!
resource policy
!
clock timezone EST 0
ip subnet-zero
!
crypto isakmp policy 1
encr 3des
authentication pre-share
group 2
lifetime 3600
crypto isakmp key cisco address 209.165.200.225
crypto isakmp key cisco address 209.165.201.1
!
crypto gdoi group group1
identity number 1
server address ipv4 209.165.201.1
server address ipv4 209.165.200.225
!
crypto map map-group1 10 gdoi
set group group1
!
interface Ethernet0/0
ip address 209.165.200.225 255.255.255.252
crypto map map-group1
!
router bgp 2000
no synchronization
bgp log-neighbor-changes
network 10.1.2.0 mask 255.255.255.0
neighbor 10.1.1.6 remote-as 5000
no auto-summary
!
ip classless
!
end
Example: Group Member 3
Group member 3 is part of a GDOI group that correlates with a VPN with which these sites are a part.
service timestamps debug datetime msec
service timestamps log datetime msec
no service password-encryption
!
hostname GM3
!
resource policy
!
clock timezone EST 0
ip subnet-zero
!
crypto isakmp policy 1
encr 3des
authentication pre-share
group 2
lifetime 3600
crypto isakmp key cisco address 209.165.200.225
crypto isakmp key cisco address 209.165.201.1
!
crypto ipsec transform-set gdoi-trans-group1 esp-3des esp-sha-hmac
crypto gdoi group group1
identity number 1
server address ipv4 209.165.200.225
server address ipv4 209.165.201.1
!
crypto map map-group1 10 gdoi
set group group1
!
interface Ethernet0/0
ip address 209.165.201.1 255.255.255.252
crypto map map-group1
!
router bgp 3000
no synchronization
bgp log-neighbor-changes
network 10.1.3.0 mask 255.255.255.0
neighbor 10.1.1.10 remote-as 5000
no auto-summary
!
ip classless
!
end
Example: Group Member 4
Group member 4 is part of a GDOI group that correlates with a VPN with which these sites are a part.
service timestamps debug datetime msec
service timestamps log datetime msec
no service password-encryption
!
hostname GM4
!
clock timezone EST 0
ip subnet-zero
!
crypto isakmp policy 1
encr 3des
authentication pre-share
group 2
lifetime 3600
crypto isakmp key cisco address 209.165.200.225
crypto isakmp key cisco address 209.165.201.1
!
crypto gdoi group group1
identity number 1
server address ipv4 209.165.200.225
server address ipv4 209.165.201.1
!
crypto map map-group1 10 gdoi
set group group1
!
interface Ethernet0/0
ip address 209.165.201.1 255.255.255.252
crypto map map-group1
!
router bgp 4000
no synchronization
bgp log-neighbor-changes
network 10.1.4.0 mask 255.255.255.0
neighbor 10.1.1.14 remote-as 5000
no auto-summary
!
ip classless
!
end
Example: Group Member 5
If a group member has multiple interfaces that are part of the same GDOI group, you should use a loopback interface to source the crypto. If a loopback interface is not used, each interface that handles encrypted traffic must register individually with the key server.
The key server sees these as separate requests and must keep multiple records for the same group member, which also means sending multiple rekeys. If crypto is sourced from the loopback interface instead, the group member registers only once with the key server.
The following configuration shows how the group member registers once with the key server:
!
interface GigabitEthernet0/1
description *** To AGG-1 ***
crypto map dgvpn
!
interface GigabitEthernet0/2
description *** To AGG-2 ***
crypto map dgvpn
!
interface Loopback0
ip address 209.165.201.1 255.255.255.255
!
crypto map dgvpn local-address Loopback0
!
Example: Passive SA
The following example displays information about crypto rules on outgoing packets:
Router# show crypto ruleset
Ethernet0/0:
59 ANY ANY DENY
11 ANY/848 ANY/848 DENY
IP ANY ANY IPSec SA Passive
IP ANY ANY IPSec Cryptomap
The following example displays the directional mode of the IPsec SA:
Router# show crypto ruleset detail
Ethernet0/0:
20000001000019 59 ANY ANY DENY -> 20000001999999
20000001000029 11 ANY/848 ANY/848 DENY -> 20000001999999
20000001000035 IP ANY ANY IPSec SA Passive
20000001000039 IP ANY ANY IPSec Cryptomap
Example: Fail-Close Mode
The following example shows that fail-close mode has been activated, and unencrypted traffic from access list 102 is allowed before the group member is registered:
crypto map map1 gdoi fail-close
match address 102
activate
crypto map map1 10 gdoi
set group ks1_group
match address 101
!
access-list 101 deny ip 10.0.1.0 0.0.0.255 10.0.1.0 0.0.0.255
access-list 102 deny tcp any eq telnet any
The following show crypto map gdoi fail-close command output shows that fail-close has been activated:
Router# show crypto map gdoi fail-close
Crypto Map: "svn"
Activate: yes
Fail-Close Access-List: (Deny = Forward In Clear, Permit = Drop)
access-list 105 deny tcp any port = 23 any
access-list 105 deny ospf any any
Additional References
Related Documents
|
|
---|---|
Cisco IOS commands (listed in an index) |
|
Cisco IOS security commands |
|
Configuring IKE and IKE policy |
|
Configuring an IPsec transform |
Standards
|
|
---|---|
No new or modified standards are supported, and support for existing standards has not been modified. |
— |
MIBs
RFCs
|
|
---|---|
RFC 2401 |
Security Architecture for the Internet Protocol |
RFC 3547 |
The Group Domain of Interpretation |
Technical Assistance
Feature Information for Cisco Group Encrypted Transport VPN
Table 3 lists the features in this module and provides links to specific configuration information.
Use Cisco Feature Navigator to find information about platform support and software image support. Cisco Feature Navigator enables you to determine which software images support a specific software release, feature set, or platform. To access Cisco Feature Navigator, go to http://www.cisco.com/go/cfn. An account on Cisco.com is not required.
Note Table 3 lists only the software release that introduced support for a given feature in a given software release train. Unless noted otherwise, subsequent releases of that software release train also support that feature.
|
|
|
---|---|---|
Cisco Group Encrypted Transport VPN |
Cisco IOS XE Release 2.3 |
Cisco Group Encrypted Transport VPN is an optimal encryption solution for large-scale IP or MPLS sites that require any-to-any connectivity with minimum convergence time, low processing, provisioning, managing, and troubleshooting overhead. The following commands were introduced or modified: address ipv4 (GDOI), clear crypto gdoi, crypto gdoi gm, debug crypto gdoi, local priority, peer address ipv4, redundancy, rekey address ipv4, rekey transport unicast, replay counter window-size, replay time window-size, sa receive-only, show crypto gdoi. |
VSA Support for GET VPN |
Cisco IOS XE Release 2.3 |
Cisco VSA (high-performance crypto engine) support was added for GDOI and GET VPN. Note This platform does not support time-based antireplay. |
GET VPN Phase 1.2 |
Cisco IOS XE Release 2.3 |
Support for time-based antireplay on the Cisco VSA was added. In Cisco IOS XE Release 2.3, this feature was implemented on the Cisco ASR 1000 series routers. |
GET VPN Enhancements |
Cisco IOS XE Release 2.3 |
These enhancements include the following features: •Fail-Close Mode This feature prevents unencrypted traffic from passing through the group member before that member is registered. The following sections provide information about this feature: –Verifying the Fail-Close Mode Status of a Crypto Map The following commands were added or modified for this feature: activate, clear crypto gdoi ks cooperative role, crypto map, match address, show crypto map. •Passive SA This feature allows a group member to be configured into passive mode permanently. The following sections provide information about this feature: The following command was added by this feature: passive. |
Glossary
DOI—Domain of Interpretation. For Internet Security Association Key Management Protocol (ISAKMP), a value in the security association (SA) payload that describes in which context the key management message is being sent (IPsec or Group Domain of Interpretation).
GDOI—Group Domain of Interpretation. For ISAKMP, a means of distributing and managing keys for groups of mutually trusted systems.
group member—Device (Cisco IOS router) that registers with a group that is controlled by the key server for purposes of communicating with other group members.
group security association—SA that is shared by all group members in a group.
IPsec—IP security. Data encryption protocol for IP packets that are defined in a set of RFCs (see IETF RFC 2401).
ISAKMP—Internet Security Association and Key Management Protocol. Protocol that provides a framework for cryptographic key management protocols.
KEK—key encryption key. Key used to protect the rekey between the key server and group members.
key server—Device (Cisco IOS router) that distributes keys and policies to group members.
MTU—maximum transmission unit. Size (in bytes) of the largest packet or frame that a given layer of a communications protocol can pass onward.
SA—security association. SA that is shared by all group members in a group.
TEK—traffic encryption key. Key that is used to protect the rekey between group members.
Appendix I: System Logging Messages
Table 4 lists GET VPN system logging (also called syslog) messages and explanations.