[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [GROW] Fwd: RFC 1104 refresh
GROW,
So this has been a topic of some debate as to whether or not it should
be in this group, but I'll go ahead and see.
Essentially, 1104 is somewhat still in tact, so I was wanting to
follow up on idea of showing the modern functionality of PBR.
As far as PBR being a routing hack to solve a local issue, I'd say
that's about mostly right. There are however notable exceptions.
Remember traffic will by default follow the IGP instead of a MPLS
tunnel. PBR is a very common way of solving this, especially on the
Juniper side of the house. Also recall PBR is the de facto standard
for routing based on the source address, which in content hosting
networks can be invaluable. Danny McPherson was also kind enough to
point out that "Some folks using flow spec in this manner."
In this light I'm essentially looking for people interested in
creating and shaping a draft of this nature.
On Mon, Apr 27, 2009 at 8:27 AM, Christopher Morrow
<christopher.morrow at gmail.com> wrote:
> On Mon, Apr 27, 2009 at 4:00 AM, Jari Arkko <jari.arkko at piuha.net> wrote:
>> I am taking the liberty of Ccing Ron and Ross/Adrian as I do believe this is
>> more operational or routing related than INT. Despite what it says further
>> down in the mail thread. I don't think I or my INT colleagues have the
>> expertise to make sure the right things are said in the document. I do
>> support the idea of publishing a document, though! Or maybe this is
>> something that GROW could tackle?
>>
>
> my initial thought is 'icky PBR!!', that aside though... I only ever
> think of PBR as a hack to solve some routing problem in a local area
> (single device, single-set-of-devices). I'm sure there are lots of new
> ways to do PBR today as opposed to when 1104 was written, but I'm not
> sure it's really a 'grow' topic.
>
> We could certainly invite David to present his topic and discuss it on
> the list though, no harm in that. It's possible there may be interest
> in helping with the doc as well from the membership.
>
> -Chris