Information About Administering the Switch
System Time and Date Management
You can manage the system time and date on your switch using automatic configuration methods (RTC and NTP), or manual configuration methods.
Note |
For complete syntax and usage information for the commands used in this section, see the Cisco IOS Configuration Fundamentals Command Referenceon Cisco.com. |
System Clock
The basis of the time service is the system clock. This clock runs from the moment the system starts up and keeps track of the date and time.
The system clock can then be set from these sources:
-
RTC
-
NTP
-
Manual configuration
The system clock can provide time to these services:
-
User show commands
-
Logging and debugging messages
The system clock keeps track of time internally based on Coordinated Universal Time (UTC), also known as Greenwich Mean Time (GMT). You can configure information about the local time zone and summer time (daylight saving time) so that the time appears correctly for the local time zone.
The system clock keeps track of whether the time is authoritative or not (that is, whether it has been set by a time source considered to be authoritative). If it is not authoritative, the time is available only for display purposes and is not redistributed.
Real Time Clock
A real-time clock (RTC) keeps track of the current time on the switch. The switch is shipped to you with RTC set to GMT time until you reconfigure clocking parameters.
The benefits of an RTC are:
-
RTC is battery-powered.
-
System time is retained during power outage and at system reboot.
The RTC and NTP clocks are integrated on the switch. When NTP is enabled, the RTC time is periodically synchronized to the NTP clock to maintain accuracy.
Network Time Protocol
The NTP is designed to time-synchronize a network of devices. NTP runs over User Datagram Protocol (UDP), which runs over IP. NTP is documented in RFC 1305.
An NTP network usually gets its time from an authoritative time source, such as a radio clock or an atomic clock attached to a time server. NTP then distributes this time across the network. NTP is extremely efficient; no more than one packet per minute is necessary to synchronize two devices to within a millisecond of one another.
NTP uses the concept of a stratum to describe how many NTP hops away a device is from an authoritative time source. A stratum 1 time server has a radio or atomic clock directly attached, a stratum 2 time server receives its time through NTP from a stratum 1 time server, and so on. A device running NTP automatically chooses as its time source the device with the lowest stratum number with which it communicates through NTP. This strategy effectively builds a self-organizing tree of NTP speakers.
NTP avoids synchronizing to a device whose time might not be accurate by never synchronizing to a device that is not synchronized. NTP also compares the time reported by several devices and does not synchronize to a device whose time is significantly different than the others, even if its stratum is lower.
The communications between devices running NTP (known as associations) are usually statically configured; each device is given the IP address of all devices with which it should form associations. Accurate timekeeping is possible by exchanging NTP messages between each pair of devices with an association. However, in a LAN environment, NTP can be configured to use IP broadcast messages instead. This alternative reduces configuration complexity because each device can simply be configured to send or receive broadcast messages. However, in that case, information flow is one-way only.
The time kept on a device is a critical resource; you should use the security features of NTP to avoid the accidental or malicious setting of an incorrect time. Two mechanisms are available: an access list-based restriction scheme and an encrypted authentication mechanism.
Cisco’s implementation of NTP does not support stratum 1 service; it is not possible to connect to a radio or atomic clock. We recommend that the time service for your network be derived from the public NTP servers available on the IP Internet.
The figure below shows a typical network example using NTP. Switch A is the NTP primary (formerly known as NTP primary), with the Switch B, C, and D configured in NTP server mode, in server association with Switch A. Switch E is configured as an NTP peer to the upstream and downstream Switch, Switch B and Switch F, respectively.
If the network is isolated from the Internet, Cisco’s implementation of NTP allows a device to act as if it is synchronized through NTP, when in fact it has learned the time by using other means. Other devices then synchronize to that device through NTP.
When multiple sources of time are available, NTP is always considered to be more authoritative. NTP time overrides the time set by any other method.
Several manufacturers include NTP software for their host systems, and a publicly available version for systems running UNIX and its various derivatives is also available. This software allows host systems to be time-synchronized as well.
NTP Stratum
NTP uses the concept of a stratum to describe how many NTP hops away a device is from an authoritative time source. A stratum 1 time server has a radio or atomic clock directly attached, a stratum 2 time server receives its time through NTP from a stratum 1 time server, and so on. A device running NTP automatically chooses as its time source the device with the lowest stratum number with which it communicates through NTP. This strategy effectively builds a self-organizing tree of NTP speakers.
NTP avoids synchronizing to a device whose time might not be accurate by never synchronizing to a device that is not synchronized. NTP also compares the time reported by several devices and does not synchronize to a device whose time is significantly different than the others, even if its stratum is lower.
NTP Associations
The communications between devices running NTP (known as associations) are usually statically configured; each device is given the IP address of all devices with which it should form associations. Accurate timekeeping is possible by exchanging NTP messages between each pair of devices with an association. However, in a LAN environment, NTP can be configured to use IP broadcast messages instead. This alternative reduces configuration complexity because each device can simply be configured to send or receive broadcast messages. However, in that case, information flow is one-way only.
NTP Security
The time kept on a device is a critical resource; you should use the security features of NTP to avoid the accidental or malicious setting of an incorrect time. Two mechanisms are available: an access list-based restriction scheme and an encrypted authentication mechanism.
NTP Implementation
Implementation of NTP does not support stratum 1 service; it is not possible to connect to a radio or atomic clock. We recommend that the time service for your network be derived from the public NTP servers available on the IP Internet.
If the network is isolated from the Internet, NTP allows a device to act as if it is synchronized through NTP, when in fact it has learned the time by using other means. Other devices then synchronize to that device through NTP.
When multiple sources of time are available, NTP is always considered to be more authoritative. NTP time overrides the time set by any other method.
Several manufacturers include NTP software for their host systems, and a publicly available version for systems running UNIX and its various derivatives is also available. This software allows host systems to be time-synchronized as well.
NTP Version 4
NTP version 4 is implemented on the switch. NTPv4 is an extension of NTP version 3. NTPv4 supports both IPv4 and IPv6 and is backward-compatible with NTPv3.
NTPv4 provides these capabilities:
-
Support for IPv6.
-
Improved security compared to NTPv3. The NTPv4 protocol provides a security framework based on public key cryptography and standard X509 certificates.
-
Automatic calculation of the time-distribution hierarchy for a network. Using specific multicast groups, NTPv4 automatically configures the hierarchy of the servers to achieve the best time accuracy for the lowest bandwidth cost. This feature leverages site-local IPv6 multicast addresses.
For details about configuring NTPv4, see the Implementing NTPv4 in IPv6 chapter of the Cisco IOS IPv6 Configuration Guide, Release 12.4T.
System Name and Prompt
You configure the system name on the Switch to identify it. By default, the system name and prompt are Switch.
If you have not configured a system prompt, the first 20 characters of the system name are used as the system prompt. A greater-than symbol [>] is appended. The prompt is updated whenever the system name changes.
For complete syntax and usage information for the commands used in this section, see the Cisco IOS Configuration Fundamentals Command Reference, Release 12.4 and the Cisco IOS IP Command Reference, Volume 2 of 3: Routing Protocols, Release 12.4.
Stack System Name and Prompt
If you are accessing a stack member through the active stack, you must use the session stack-member-number privileged EXEC command. The stack member number range is from 1 through 8. When you use this command, the stack member number is appended to the system prompt. For example, Switch-2# is the prompt in privileged EXEC mode for stack member 2, and the system prompt for the switch stack is Switch.
Default System Name and Prompt Configuration
The default switch system name and prompt is Switch.
DNS
The DNS protocol controls the Domain Name System (DNS), a distributed database with which you can map hostnames to IP addresses. When you configure DNS on your switch, you can substitute the hostname for the IP address with all IP commands, such as ping , telnet , connect , and related Telnet support operations.
IP defines a hierarchical naming scheme that allows a device to be identified by its location or domain. Domain names are pieced together with periods (.) as the delimiting characters. For example, Cisco Systems is a commercial organization that IP identifies by a com domain name, so its domain name is cisco.com. A specific device in this domain, for example, the File Transfer Protocol (FTP) system is identified as ftp.cisco.com.
To keep track of domain names, IP has defined the concept of a domain name server, which holds a cache (or database) of names mapped to IP addresses. To map domain names to IP addresses, you must first identify the hostnames, specify the name server that is present on your network, and enable the DNS.
Default DNS Settings
Feature |
Default Setting |
---|---|
DNS enable state |
Enabled. |
DNS default domain name |
None configured. |
DNS servers |
No name server addresses are configured. |
Login Banners
You can configure a message-of-the-day (MOTD) and a login banner. The MOTD banner is displayed on all connected terminals at login and is useful for sending messages that affect all network users (such as impending system shutdowns).
The login banner is also displayed on all connected terminals. It appears after the MOTD banner and before the login prompts.
Note |
For complete syntax and usage information for the commands used in this section, see the Cisco IOS Configuration Fundamentals Command Reference, Release 12.4. |
Default Banner Configuration
The MOTD and login banners are not configured.
MAC Address Table
The MAC address table contains address information that the switch uses to forward traffic between ports. All MAC addresses in the address table are associated with one or more ports. The address table includes these types of addresses:
-
Dynamic address—A source MAC address that the switch learns and then ages when it is not in use.
-
Static address—A manually entered unicast address that does not age and that is not lost when the switch resets.
The address table lists the destination MAC address, the associated VLAN ID, and port number associated with the address and the type (static or dynamic).
Note |
For complete syntax and usage information for the commands used in this section, see the command reference for this release. |
MAC Address Table Creation
With multiple MAC addresses supported on all ports, you can connect any port on the switch to other network devices. The switch provides dynamic addressing by learning the source address of packets it receives on each port and adding the address and its associated port number to the address table. As devices are added or removed from the network, the switch updates the address table, adding new dynamic addresses and aging out those that are not in use.
The aging interval is globally configured. However, the switch maintains an address table for each VLAN, and STP can accelerate the aging interval on a per-VLAN basis.
The switch sends packets between any combination of ports, based on the destination address of the received packet. Using the MAC address table, the switch forwards the packet only to the port associated with the destination address. If the destination address is on the port that sent the packet, the packet is filtered and not forwarded. The switch always uses the store-and-forward method: complete packets are stored and checked for errors before transmission.
MAC Addresses and VLANs
All addresses are associated with a VLAN. An address can exist in more than one VLAN and have different destinations in each. Unicast addresses, for example, could be forwarded to port 1 in VLAN 1 and ports 9, 10, and 1 in VLAN 5.
Each VLAN maintains its own logical address table. A known address in one VLAN is unknown in another until it is learned or statically associated with a port in the other VLAN.
When private VLANs are configured, address learning depends on the type of MAC address:
-
Dynamic MAC addresses learned in one VLAN of a private VLAN are replicated in the associated VLANs. For example, a MAC address learned in a private-VLAN secondary VLAN is replicated in the primary VLAN.
-
Static MAC addresses configured in a primary or secondary VLAN are not replicated in the associated VLANs. When you configure a static MAC address in a private VLAN primary or secondary VLAN, you should also configure the same static MAC address in all associated VLANs.
MAC Addresses and Switch Stacks
The MAC address tables on all stack members are synchronized. At any given time, each stack member has the same copy of the address tables for each VLAN. When an address ages out, the address is removed from the address tables on all stack members. When a Switch joins a switch stack, that Switch receives the addresses for each VLAN learned on the other stack members. When a stack member leaves the switch stack, the remaining stack members age out or remove all addresses learned by the former stack member.
Default MAC Address Table Settings
The following table shows the default settings for the MAC address table.
Feature |
Default Setting |
---|---|
Aging time |
300 seconds |
Dynamic addresses |
Automatically learned |
Static addresses |
None configured |
ARP Table Management
To communicate with a device (over Ethernet, for example), the software first must learn the 48-bit MAC address or the local data link address of that device. The process of learning the local data link address from an IP address is called address resolution .
The Address Resolution Protocol (ARP) associates a host IP address with the corresponding media or MAC addresses and the VLAN ID. Using an IP address, ARP finds the associated MAC address. When a MAC address is found, the IP-MAC address association is stored in an ARP cache for rapid retrieval. Then the IP datagram is encapsulated in a link-layer frame and sent over the network. Encapsulation of IP datagrams and ARP requests and replies on IEEE 802 networks other than Ethernet is specified by the Subnetwork Access Protocol (SNAP). By default, standard Ethernet-style ARP encapsulation (represented by the arpa keyword) is enabled on the IP interface.
ARP entries added manually to the table do not age and must be manually removed.
For CLI procedures, see the Cisco IOS Release 12.4 documentation on Cisco.com.