Access List-Based RBSCP

The Access List-Based Rate-Based Satellite Control Protocol (RBSCP) feature provides the following benefits:

• It allows you to selectively apply the TCP ACK splitting feature of RBSCP to any outgoing interface. TCP ACK splitting is a benefit because it reduces the effect of long latencies characteristic of satellite links. Applying this feature selectively by using an access list is a benefit because you control which packets are subject to TCP ACK splitting.
• It has no tunneling or queueing overhead that is associated with RBSCP tunnels.
• It provides more interoperability with other Cisco IOS features (such as TCP/IP header compresssion, DMVPN, and QoS) because the TCP and Stream Control Transmission Protocol (SCTP) packets are no longer encapsulated with an RBSCP/IP header.
• This feature works on process switched forwarding, fast switching, or CEF.
• It preserves the internet end-to-end principle.

Rate-Based Satellite Control Protocol (RBSCP) was designed for wireless or long-distance delay links with high error rates, such as satellite links. RBSCP can improve the performance of certain IP protocols, such as TCP and IP Security (IPsec), over satellite links without breaking the end-to-end model. For instructions on how to implement RBSCP over a tunnel, see the “Implementing Tunnels” chapter of the Interface and Hardware Component Configuration Guide .

The TCP ACK splitting capability of RBSCP can be implemented without a tunnel, by using an IP access list, as shown in the figure below. The TCP ACK splitting occurs at the outgoing interface between the router and the internal network or Internet. It does not occur over the link to the satellite.

TCP ACK splitting is a software technique to improve performance for clear-text TCP traffic using acknowledgment (ACK) splitting, in which a number of additional TCP ACKs are generated for each TCP ACK received. TCP ACK splitting causes TCP to open the congestion window more quickly than usual, thus decreasing the effect of long latencies. TCP will generally open the congestion window by one maximum transmission unit (MTU) for each TCP ACK received. Opening the congestion window results in increased bandwidth becoming available. Configure this feature only when the satellite link is not using all the available bandwidth. Encrypted traffic cannot use TCP ACK splitting.

The size argument in the ip rbscp ack-splitcommand determines how many TCP ACKs are generated from the incoming TCP ACK, as shown in the figure below.

If n ACKs are configured and M is the cumulative ACK point of the original TCP ACK, the resulting TCP ACKs exiting the router will have the following cumulative ACK points:

M-n+1, M-n+2, M-n+3,...M

For example, if the size argument is set to 5, and the access list permits a TCP ACK with a cumulative ACK acknowledging bytes to 1000, then the resulting TCP ACKs exiting the router will have the following cumulative ACK points:

TCP ACK (996) (1000-5+1)

TCP ACK (997) (1000-5+2)

TCP ACK (998) (1000-5+3)

TCP ACK (999) (1000-5+4)

TCP ACK (1000) (1000-5+5)

The Access List-Based RBSCP feature will accept a numbered or named, standard or extended IP access list. The access list controls which packets are subject to TCP ACK splitting. That is, the feature is applied to packets that a permit statement allows; the feature is not applied to packets that a deny statement filters.

An instance of this feature consists of an access list and an ACK split value. An ACK split value of 0 or 1 indicates that this feature is disabled (that is, no ACK split will be done). The ACK split value range is 0 through 32.

An interface can use only one instance of this feature at a time. Each instance of this feature can be used on multiple interfaces.

If you configure this feature but it refers to a nonexistent access list, this is interpreted as having an access list that denies all traffic from being processed by the access list-based RBSCP feature, so the feature is essentially disabled and the traffic goes through the normal switching path.

If both an RBSCP tunnel and an instance of the Access List-Based RBSCP feature are enabled along a routing or switching path, the TCP ACKs detunneled from the RBSCP tunnel will be ACK split according to the tunnel configuration and the Access List-Based RBSCP split parameters on the outgoing interface are effectively disabled.

SUMMARY STEPS

1.    enable

2.    configure terminal

3.    interface type number

4.    ip rbscp ack-split size {access-list-name | access-list-number} out

5.    Although it is not required, you should repeat this task on the router that is on the other side of the satellite, on the outgoing interface facing the network, not the satellite. Use a different access list.

DETAILED STEPS

	Command or Action	Purpose
Step 1	enable Example: Router> enable	Enables privileged EXEC mode. Enter your password if prompted.
Step 2	configure terminal Example: Router# configure terminal	Enters global configuration mode.
Step 3	interface type number Example: Router(config)# interface ethernet 1	Specifies an interface. Specify an interface that is facing your internal network, opposite the satellite network.
Step 4	ip rbscp ack-split size {access-list-name | access-list-number} out Example: Router(config-if)# ip rbscp ack-split 6 101 out	Configures RBSCP on the outgoing interface for packets that are permitted by the specified access list. The ACK split size determines the number of ACKs to send for every ACK received. An ACK split value of 0 or 1 indicates that this feature is disabled (that is, no ACK split will be done). The range is 0 through 32. See "TCP ACK Splitting". In this example, access list 101 determines which packets are subject to TCP ACK splitting.
Step 5	Although it is not required, you should repeat this task on the router that is on the other side of the satellite, on the outgoing interface facing the network, not the satellite. Use a different access list.	--