Intersite Rate Limiting per service/connection/customer.

[steve.wadge]

I have a scenario where customers have dedicated connections/infrastructure into a hosted environment at 2 x Data Centres (1 x primary, 1 x backup) with shared bandwidth between Data Centres.

I wish to limit/restrict the amount of bandwidth between data centres on a per customer basis i.e. allocate say 2Mb per customer to stop any single customer grabbing all the bandwidth.

The intersite connectivity is between 6509's with 1 x LES100 primary & 2 x 45Mbs ATM secondary. The customers circuits are generally terminated on 2600's with FastEthernet connectivity between 2600's & 6500's.

The question is how to/the best way of limiting this traffic. As this is a relatively new requirement all options are open (including re-engineering). I assume that either some traffic-shaping, class based shaping, rate limiting will need enabling but which is most appropriate & on what type of interface.

[vkapoor5]

You can use CBWFQ along with CAR to limit.

Please refer the following link for more information....

http://www.cisco.com/en/US/products/hw/routers/ps368/products_module_configuration_guide_chapter09186a008007cb72.html

[jishi]

Can you please descript the whole setup in more detail? How do the customers connect to the data center? What type of traffic you want to limit? Only SNA or all the traffic?

If the customers come into the data center over different PVC, and if you want to limit all traffic, you can do ATM policing on the central site or ATM traffice shaping on the remote customer site, if you have control over there.

Jing

[steve.wadge]

Jing,

Each client has 2 x leased line connections (1 into each of 2 data centres). The Data centre's have ATM - ds3 connections between them. Each datacentre houses a hosted environment for each client which is housed on a dedicated Lan per client.

The hosted enironment works in an active standby configuration so traffic should only be traversing one of the client leased lines at any one time.

The specific issue is how to limit/restrict the amount of traffic between data centre's that each client could possibly use i.e in the case of database replication. Ideally I'd like to be able to restrict this to the same bandwidth as there leased line although as long as I can impose a limit to potentially stop a single clients traffic overloading the inter-Data Centre ATM links.

Regards,

Steve.