SIP in higher education

[(Dennis Baron)]

SIP.edu Conference Call February 16, 2006

*Attendees*

Dennis Baron, MIT
John Covert, Independent Consultant
Arthur Gaylord, Woods Hole Oceanographic Institute
Candace Holman, Columbia
Jerry Keith, UC Riverside
Jeff Kuure, Internet 2
Christian Schlatter, UNC
Ben Teitelbaum, Internet 2
John Todd, Tello
Chris Trown, University of Oregon
Joe Mailander, University of Oregon
Chris Trown, University of Oregon
Dave Zimmerman, Berkeley

*Discussion*

Today's call begins with Dave Zimmerman talking about a recent
RFP at Berkeley for a new PBX. The RFP specifically mentions SIP
compatibility. Dave is in the network group, and finds more voice related
technology being directed towards his department. He's interested in
finding some resources that might tell him what sort of equipment other
institutions have used for SIP gateways. Dennis points him towards the
SIP.edu cookbook on MIT's website, which contains deployment overviews
provided by several institutions. It is not completely up to date,
but should provide some information. Dennis also mentions the SIP.edu
email list as a source, as participants would be likely to provide more
current information.

Chris Trown is in a similar situation. Recently, Avaya has announced
the end-of-life for some products sooner than they had expected, so
they will be looking for new equipment. He's interested in OpenSER and
Asterisk, but feels that the lack of management tools, particularly for
non-technical users, is a major issue. He has investigated SIP Foundry,
but has run into some problems and missing features.

Dennis asks if it would be beneficial to have someone on a future call to
discuss PBX replacement with a focus on SIP interoperability. Participants
would find this interesting, but several comment that they would have
trouble finding someone to do this - most universities are still tied
to legacy PBXs. John Covert mentions that people are getting into using
Asterisk and PRI span for VOIP, like most people on this call. There are
a lot of hybrid deployments, and most PBX vendors that do VOIP use H.323
and other proprietary protocols. Dennis asks if the person who wrote
Berkeley's RFP be interested in participating in a future call as there
seems to be questions about replacing existing PBXs with a future plans
for doing VOIP. People would like to know what needs to be done in terms
of call waiting and management. Dave will talk to someone on the committee
about participating in a future call. Dennis will investigate user
management of open source SIP software for a possible future discussion.

Christian Schlatter mentions that SIP PBX is sort of a misnomer as most
of the power of SIP lies in the endpoints. Most vendors don't want to
supply such a thing because it goes against their revenue stream, which
relies on centralized control.

Following this is a discussion about SIP handsets, specifically
the solicitation of feedback as to what products are worth
investigating. Polycom phones were recommended, and the Cisco ATA 186
and 188 models.

JohnC mentioned the Cisco Linksys SPA 941/942. They use the
Sipura firmware which is more advanced than the Cisco 7900 firmware. They
offer SRTP, TLS, encryption, and line presence, but the require a pretty
advanced SIP infrastructure to take advantage of all the features.

Dennis has converted from his obsolete Pingtel phone to a Snom 360,
which works fine and has a lot more features than the Cisco phones. There
are a few oddities, such as poor documentation. These are being proposed
for some SIP presence experiments on campus. Dennis found the SIP
functionality of the Polycom 601 to be very limited. It only offered only
one global outbound proxy, for example. The speakerphone quality and display
were nicer on the Polycom, but Dennis is back to the Snom.

Christian Schlatter asks about firewall and NAT traversal in the Snom
phones.  Dennis has not personally played with this, but he mentions
that the frequent software updates for the Snom phones cause support of
some things to be rather inconsistent.  But they are supposed to be
supporting STUN and ICE - and had support for UPnP at one point.

Following this is a discussion of hard phone deployment on campuses. Do
evaluations of hardware consider faculty only, or are student dorms
considered as well? Most agree that they don't want to put expensive
phones into dorm rooms. There are also wiring, power, and air conditioning
issues to consider for dormitory deployment.  Converting buildings to
SIP is often prohibitively expensive, and is therefore reserved for new
or remodeled buildings.

Art Gaylord joins the call, and Dennis asks about WHOI's hard phone
deployment. WHOI has 200 Snom phones in use, with 500 more ready to
go. They are rolling out 10-15 per day. He believes that in 3 to 5 years
most people won't be concerned with having a desk phone - a combination
of desktop PC soft phones and mobile Wi-Fi phones will supplant them. He
is asked about firmware upgrades and remote debugging of for so many
phones. This is accomplished with some custom software for initial
deployment and a web interface to update deployed phones.

Dennis moves on to the ISN initiative. He, Ben, John Todd, and Mark
Silis talked earlier in the week, and are moving along with DNS
infrastructure. DNS entry management has also being discussed, and a
mailing list has been set up. Instructions for joining are available on
the freenum.org website. There is a low-volume, one-way announcement list,
as well as a general discussion list for questions and assistance. There
are currently about 50 organizations who have applied for ITAD numbers,
including Apple - and Google has expressed interest. Any educational
institutions who have questions can talk to John Todd for assistance.

Dennis is asked about the issues MIT has had with their 5ESS. They have
had problems making ISN calls, and are now building a separate trunk group
and will look into doing digit collection on the Cisco gateway. If that
does not work, they will look into doing it with an Asterisk box. They
will probably end up with two trunk groups - one that will allow dialing
a variable number digits with a * which will be cut through, and another
for a variable number of digits, such as the subscriber portion, which
will not be cut through and would allow call forwarding.

Finally, Candace poses a question. In the Presence and Integrated
Communications (PIC) working group for Internet2, they have been discussing
XMPP.  She's had been reluctant to change her mind from SIP to XMPP, but
has read more of the Jabber standards and has changed her opinion. She
was not around during the decision to shift from H.323 to SIP in the VOIP
working group, and wonders if this is the same thing. She's interested
in the historical perspective on the shift from H.323 to SIP, and if
another shift from SIP to XMPP could be expected in the next year or two.

Ben does not know if this a case of history repeating itself, but would
be interested in knowing the history of the previous shift. Dennis
asks if there is anyone who would be able to provide an overview of the
history. He wonders if he was a newly introduced to these protocols, if he
would bother with SIP and would rather investigate XMPP, which is leaning
towards becoming more SIP-like with audio and video support. This leads
to a discussion on cultural differences and support for various protocols.

The next call will be on March 2nd.