> Is there any major showstopper bug in wccp v1? We have a Cisco box that
> for now we won't be able to upgrade to wccp v2 and wanted to know if
> attaching a Cacheflow to it via wccp v1 would be problematic.
no "showstoppers" in IOS for WCCPv1 per-se (although you neglect to
mention what model router and what release of IOS, so ....), but ..
there's some newly-posted information on CCO about WCCP.
see http://www.cisco.com/warp/public/732/wccp/
included in this (besides some broken links that need to be fixed)
are some powerpoint presentations that go thru what releases have
WCCP in what switching paths, network architectures and features
that may be useful.
while this talks about WCCP, it does NOT talk about whether CacheFlow's
implementation is complete or bug-proof. i don't think it would be
politically correct for me to attempt to answer that in a public forum --
probably best you talk to cacheflow about that.
as some general statements:
- if you're on 72xx/75xx, try to stay with the 12.0S train. you
have WCCP both in CEF/dCEF (as appropriate) and some very neat
traffic-engineering features (using bgp-policy-propogation for
intercept/bypass decisions)
- the packet-forwarding mechanisms for WCCPv1 and WCCPv2 are
practically identical. the only exception is the newer Catalyst 6x00
forwarding path where encapsulation-method and hash-method are
negotiated depending on router/cache/switch capabilities and
preferences.
- Catalyst 6x00 will "scale" higher than any other platforms (with
other platform that have h/w-based forwarding) to follow.
latest Catalyst 6500+MSFC2/PFC2 combinations can do WCCP[v2] at 30M PPS.
cheers,
lincoln.
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:22 EDT