RE: [EFM] Is "campus" P2MP out of scope?
In many cases the campus environment is very important as an access
E.g. Universities, Army, city offices in a city carrier environment ...
In those scenarios the network is difference from business access since it
has to deal with
different kind of application, QoS and security issues.
e.g. In an enterprise environment the inside traffic security is less
important than in a campus environment.
Thus I think situation "c" is important and it has different demands than
an enterprise environment.
[mailto:firstname.lastname@example.org]On Behalf Of Roy Bynum
Sent: Thursday, November 29, 2001 7:04 AM
To: glen.kramer@xxxxxxxxxxxx; millardo@xxxxxxxxxxxxxxxxxx
Subject: RE: [EFM] Is "campus" P2MP out of scope?
Service providers bring the customers' traffic back to an access point that
becomes the revenue generation point that creates the billable
revenue. Having been with a service provider for over 10 years, I have
never seen service deployment that you described in situation
"c". Situation "c" is more of an "enterprise" type of deployment not a
"commercial" one. I believe that situation "c" is out of scope of this TF.
At 12:06 PM 11/26/2001 -0800, glen.kramer@xxxxxxxxxxxx wrote:
>You are right that both enterprise and campus networks (which are LANs) are
>out of scope of this TF.
>I however, want to clarify in what context they were mentioned on the
>EFM is charged with defining "access network". But when we talk about
>functional requirements of access networks, we realize that different
>applications have different requirements.
>a. Residential access network - not much traffic from user to user (ONU to
>ONU), but downstream broadcasting is important (video broadcasting). And
>thus it was stated that a combination of point-to-point and shared
>(P2P+SE) makes sense.
>b. Business access network (what sometimes referred to as enterprise access
>network) - no downstream broadcasting video is needed, and so downstream
>broadcasting is not important. Also not much traffic between ONUs. This is
>the case for P2P emulation only.
>c. Campus access network - the difference from business access is that all
>tailend nodes belong to the same administrative domain. There can be a fair
>amount of out-bound traffic as well as ONU-to-ONU traffic. There is a fair
>number of lowtech campuses that don't have or don't want to have their own
>IT department to maintain a campus network. This is a good place for P2MP
>network with a shared emulation. It is still an access network, but better
>optimized for ONU-to-ONU traffic.
>It is not up to the standard to decide in what environment the access
>network is to be used. But, standard can allow multitude of configurations
>(as it does for LANs) that each vendor will make decision on.
> > Howard,
> > I am seeing several references to a "enterprise" type of "campus"
> > deployment as a target for P2MP optical services. I may be
> > mistaken, but I
> > thought that this TF was working on support of "subscription
> > networks"
> > which, by my understanding, are commercial service access
> > networks, not
> > enterprise networks. Am I mistaken? If I am not, then that
> > would make the
> > need to support enterprise campus networks somewhat out of scope.
> > I hate to see a lot of effort put into trying to support
> > campus networks
> > for ubiquitous shared access over optical media. From my experience,
> > ubiquitous shared networks have an effective utilization of
> > about 30%,
> > depending on the number of nodes on the media. The support and
> > maintenance of that type of topology in the enterprise campus
> > environment
> > would be very similar to the old "coax" system of years ago.
> > At the lower
> > utilization, an the high maintenance labor costs, the higher
> > cost of the
> > optical media would not be cost effective. I don't see much
> > of a market
> > for that type of deployment.
> > Thank you,
> > Roy Bynum