Application Acceleration
Reply
Visitor
JoTraGo
Posts: 3
Registered: ‎08-17-2009
0

Outbound QOS max limit in relation to inbound bandwidth at a target site

A typical scenario with HQ feeding an MPLS cloud with multiple branches.

HQ has 6Mb/s link to the cloud, Branches have 1Mb/s link to the cloud

HQ is oversubscribed with Bandwidth detection enabled and Link speeds for remote sites are set accurately.

Lets say that the Business standard class is set for a Max Bandwidth of 50%

For any given branch WIll Business Standard be limited to 3Mb/s (50%of HQ Link) or 1Mb/s (50% of Remote Link)?

 

 

In a similar but slightly more complex scenario there are TWO Data Centres, one at HQ as above and one hosted at an Outsource Service Provider (OSP) each carrying different traffic for different services. We need to ensure that the inbound traffic for any given branch is not saturated.

If Bandwidth detection is enabled will the HQ & OSP WX be able automatically manage their outbound bandwidth? Or would I need to manually set some limits, eg50% each to ensure that the inbound link for any branch is not overloaded?

Visitor
NRamesh
Posts: 9
Registered: ‎08-15-2011
0

Re: Outbound QOS max limit in relation to inbound bandwidth at a target site

Answer to 1st queries

Bandwidth detection detects 1Mb/s for remote link then it will allow 50% (0.5M) to Remote Link

Example: If Bandwidth detection detects 800K then it will allow 400K to remote link.

 

Answer to 2nd queries

Only the way is to custom the QoS settings for OSP endpoint as configure min BW as 'zero' on HQ WX, So that during the WAN congestion the outbound BW will be restricted on HQ for OSP, During this situation TCP will slow down/control the flow between the client to server communication.

 

-Ramesh

Copyright© 1999-2013 Juniper Networks, Inc. All rights reserved.