[a-nsp] VLAN in SNMP Interface Table

Mon Dec 9 19:47:59 EST 2019

Mike.  Arista 7280R/R2 and 7500R/R2 series can support per port/vlan
counters via Layer 2 subinterface (packets, bytes and rates).  The feature
TOI is here (the egress counter restriction is resolved in the latest EOS
release):
https://eos.arista.com/eos-4-21-1f/l2-sub-interfaces/

L2 subinterfaces are a different config model than a trunk interface
(benefiting from things like counters among others).  There are also scale
and feature interaction considerations (see TOI).  The counters are exposed
via SNMP as well.

There is also the ability to match vlan as part of a policy map for which
class matches can be counted (but this is ingress only with packet wise
counting):
https://eos.arista.com/eos-4-17-0f/per-port-per-vlan-qos/

There is of course sFlow for which many platforms support
aggressive sampling rates which may provide the granularity you require.

Regards.

david

On Mon, Dec 9, 2019 at 2:09 PM Patrik Olsson via arista-nsp <
arista-nsp at puck.nether.net> wrote:

>
>
>
> ---------- Forwarded message ----------
> From: Patrik Olsson <darkiesan at me.com>
> To: Mike Hammett <arista-nsp at ics-il.net>, Doug McIntyre <merlyn at geeks.org>,
> Patrik Olsson via arista-nsp <arista-nsp at puck.nether.net>
> Cc:
> Bcc:
> Date: Mon, 9 Dec 2019 20:09:24 +0100
> Subject: Re: [a-nsp] VLAN in SNMP Interface Table
> Counting packets gives no insight, not even per VLAN.
> I would rather get a good free (or cheap) sFlow probe, like InMons.
> There you can get graphs in anyway you like it. Per port, Per VLAN etc.
> And it also gives graphs and insight to application layer (i.e. L3 and L4
> headers in the packet).
> And sFlow is supported in Arista gear.
>
> Just my 50 cent
> Patrik
>
>
> On 9 Dec 2019, at 15:35, Mike Hammett <arista-nsp at ics-il.net> wrote:
>
> Except VLANs are touched by the layer 2 switch and routed accordingly.
> Using the mail\post office analogy, it would be like a suite or apartment
> number at a given address.
>
>
> I'm not going to try to force Arista into my need. If Arista doesn't do
> it, then I'll move on. It doesn't seem like they do.
>
>
>
> -----
> Mike Hammett
> Intelligent Computing Solutions <http://www.ics-il.com/>
> <https://www.facebook.com/ICSIL>
> <https://plus.google.com/+IntelligentComputingSolutionsDeKalb>
> <https://www.linkedin.com/company/intelligent-computing-solutions>
> <https://twitter.com/ICSIL>
> Midwest Internet Exchange <http://www.midwest-ix.com/>
> <https://www.facebook.com/mdwestix>
> <https://www.linkedin.com/company/midwest-internet-exchange>
> <https://twitter.com/mdwestix>
> The Brothers WISP <http://www.thebrotherswisp.com/>
> <https://www.facebook.com/thebrotherswisp>
> <https://www.youtube.com/channel/UCXSdfxQv7SpoRQYNyLwntZg>
> ------------------------------
> *From: *"Doug McIntyre" <merlyn at geeks.org>
> *To: *arista-nsp at puck.nether.net
> *Sent: *Monday, December 9, 2019 8:32:07 AM
> *Subject: *Re: [a-nsp] VLAN in SNMP Interface Table
>
>
> I'd count the packets exiting a port as doing some work.
>
> Maybe a better analogy would be you have a bunch of envelopes sent
> through the post office. You are asking a counter to see how many
> pieces of paper are in each envelope. The post office only delivers
> the envelopes, so we have a total count of envelopes. Only if we need
> to burst open the envelopes (layer-3) and start routing the pieces of paper
> based on their top tag, will we have counters on them. As long as the
> post office keeps forwarding intact envelopes (ie. layer-2), we won't have
> visibility into inside.
>
>
>
>
> On Mon, Dec 09, 2019 at 07:40:00AM -0600, Mike Hammett wrote:
> > If the first paragraph were true, we'd never have port counters on a
> switch.
> >
> >
> >
> >
> > -----
> > Mike Hammett
> > Intelligent Computing Solutions
> >
> > Midwest Internet Exchange
> >
> > The Brothers WISP
> >
> > ----- Original Message -----
> >
> > From: "Doug McIntyre" <merlyn at geeks.org>
> > To: arista-nsp at puck.nether.net
> > Sent: Monday, December 9, 2019 7:36:05 AM
> > Subject: Re: [a-nsp] VLAN in SNMP Interface Table
> >
> > On Sun, Dec 08, 2019 at 06:02:47PM -0600, Mike Hammett wrote:
> > > This looks like it's only on layer 3 interfaces, which doesn't do me a
> whole lot of good.
> >
> >
> > I don't know of any hardware that keeps track of layer-2 VLAN packet
> > counts going through the switch. Generally, stats are updated on times
> > when switches have to _do_ something. Forwarding along layer-2 is
> > usually done in ASICs and the switch doesn't do anything special for
> > decoding the VLAN headers. OOTH, doing a layer-3 routing termination
> > is doing something, which is where I measure everything, on the
> > layer-3 edge. And every type of hardware I touch can monitor that.
> >
> > As posted before, if you need this sort of thing, you'll probably have
> > to go to sflow type collectors in order to get what you want. Or
> > consider making your network do layer-3 VLAN terminations somewhere so
> > that you can measure it.
> >
> > Perhaps there is some weird EVPN/vxlan setup you could make that the
> > switch will have to do punting of traffic out to a VTEP, and you could
> > measure those counts, but I have zero experience monitoring such a setup.
>
> >
> > --
> > arista-nsp mailing list
> > arista-nsp at puck.nether.net
> > https://puck.nether.net/mailman/listinfo/arista-nsp
> >
> --
> arista-nsp mailing list
> arista-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/arista-nsp
>
> --
> arista-nsp mailing list
> arista-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/arista-nsp
>
>
>
>
>
> ---------- Forwarded message ----------
> From: Patrik Olsson via arista-nsp <arista-nsp at puck.nether.net>
> To: Mike Hammett <arista-nsp at ics-il.net>, Doug McIntyre <merlyn at geeks.org>,
> Patrik Olsson via arista-nsp <arista-nsp at puck.nether.net>
> Cc:
> Bcc:
> Date: Mon, 9 Dec 2019 20:09:24 +0100
> Subject: Re: [a-nsp] VLAN in SNMP Interface Table
> --
> arista-nsp mailing list
> arista-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/arista-nsp
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://puck.nether.net/pipermail/arista-nsp/attachments/20191209/60137a2c/attachment.htm>