nanog mailing list archives
RE: Overflow circuit
From: "Michel Py" <michel () arneill-py sacramento ca us>
Date: Fri, 26 Mar 2004 15:46:50 -0800
Joe wrote: I am looking for advice on technique or products that can solve the following challenge ... Two private line T1's between A and B - one terrestial T1 with >200 ms RTT, the other T1 is over satellite with ~500 ms RTT. The circuits are being used for mixed VoIP (70%) and data (30%) applications. To achieve optimal voice quality, we want to route all VoIP calls over the terrestrial T1 until it is "full", then divert all subsequent VoIP calls over the satellite T1 (**while existing VoIP calls continue to be routed over the terrestial T1).
I have a cheesy solution..... 1. You need a routing protocol that load-balances. I have done it before with EIGRP, you mileage may vary with others IGPs. 2. Cook a custom metric that will prefer the terrestrial link until full (a mix of delay and load will likely do). 3. Use policy routing to send anything but voice to the satellite link. There's a catch here: even if there is no voice, your data is limited to satellite, which works with your traffic pattern. Your data is forced to the sat, your voice picks the best path as per your custom metric. 4. On the satellite link, prioritize voice over data. 5. Fine-tune the variance. Michel.
Current thread:
- Re: Overflow circuit, (continued)
- Re: Overflow circuit Patrick Murphy (Mar 26)
- Re: Overflow circuit Alexei Roudnev (Mar 26)
- RE: Overflow circuit Mailing List Subscriptions (Mar 27)
- Re: Overflow circuit Alex Rubenstein (Mar 27)
- Re: Overflow circuit Patrick Murphy (Mar 27)
- Re: Overflow circuit alex (Mar 28)
- RE: Overflow circuit Mailing List Subscriptions (Mar 27)
- Re: Overflow circuit Alexei Roudnev (Mar 26)
- Re: Overflow circuit Stephen Sprunk (Mar 28)
- RE: Overflow circuit Joe Chin (Mar 28)
- Re: Overflow circuit Alexei Roudnev (Mar 28)
- RE: Overflow circuit Joe Chin (Mar 28)
- RE: Overflow circuit Michel Py (Mar 26)
- Re: Overflow circuit W.D.McKinney (Mar 27)
- RE: Overflow circuit Michel Py (Mar 27)
- Re: Overflow circuit Alexei Roudnev (Mar 27)
- Re: Overflow circuit Steven M. Bellovin (Mar 27)
- Re: Overflow circuit Alexei Roudnev (Mar 27)
- Re: Overflow circuit Steven M. Bellovin (Mar 27)
- Re: Overflow circuit Rafi Sadowsky (Mar 27)
- Re: Overflow circuit Alexei Roudnev (Mar 27)
- RE: Overflow circuit Mailing List Subscriptions (Mar 28)
- Re: Overflow circuit Alexei Roudnev (Mar 27)
- Re: Overflow circuit Patrick Murphy (Mar 26)