The Pulver Report: May 15, 2000

THE PULVER REPORT™ : The May 15, 2000 Issue:

In this Issue:

• Heard on the Net
• Jeff's Change Agent of the Month: The RIM 950 and GoAmerica.net
• Feedback from April 17th issue: ATM/MPLS clarification
• DSL needs a bit of PacketCable
• 5th SIP Bakeoff - August 8-10 in Melville, NY
• Trends in Instant Messaging: Feedback from our College Intern
• Instant Messaging 2000 - May 23-25, Boston, MA
• The Pluses and Minuses of Research and Development.

Heard on the Net
- People on the Move:

Dean Willis, Robert Sparks, Chris Cunningham and Ben Campbell have all left MCI Worldcom and have joined dynamicsoft.

Ken McInerney has left Salix (Tellabs) to join BroadSoft as Director, Product Management.

Gary Bean has left Salix (Tellabs) to join BroadSoft as Director, Strategic Accounts.

Zig Serafin, former SVP Callware, most recently Business Planning lead in Windows Division of Microsoft now a lead for Internet Telephony strategy in Microsoft's M&A/Corporate Development arm.

Curt Stamford has left Lucent Technologies.

Jim Cassella has left Lucent Technologies to accept a sales position with Clarent Corporation.

Hernan Popper has left his position as COO at Keycom Argentina, a player in Value Added Services in Argentina, and has joined COMSAT Argentina as VOIP/Telephony Product Manager for the deployment of the Long Distance and International Telephony services.

Mark Vitale has left Telcorida and joined Lucent Technologies' Network Care Professional Services team.

(Please email: people@pulver.com to report a change in your position.)

- Companies on the Move:

The VON Job marketplace continues to be hot. A record 38 jobs are currently posted at ( http://pulver.com/jobs ).

Acquisitions: Clarent Corporation recently signed an agreement to acquire ACT Networks, Inc. for a deal valued at $189 million.

A number of companies have recently announced funding activities. Some of those reporting include:

Dialpad.com has secured a $16.75 million initial round of venture funding led by CMGI @Ventures. Additional investors include Sterling Payot Company and Serome Technology Inc.

Tundo Corporation has secured $30 million in equity and debt financing. New investors included JK&B Capital, Goldman Sachs, Comverse, and Newbury Ventures.

IPVoice.com has secured $2.5 million equity financing from institutional investment groups.

Recent stock market performance has slowed down the pace of VON S1 filings. Since the introduction of the pulver.com VON Index(tm) set to 100 on March 28th, the index ended April, 2000 at 69.94. The VON Index has seen intraday high/lows ranging between 80.02 and 64.15 during the first half of May, 2000.

For details regarding the pulver.com VON Index, please visit ( http://pulver.com/vonindex ). For up-to-date daily Industry news, please feel free to visit (http://pulver.com/news ). If you have news you would like to see posted, please email: editor@pulver.com. To schedule time as a a possible guest on pulverradio.com (http://www.pulverradio.com ), please email: matt@pulver.com.

Jeff's Change Agent of the Month: The RIM 950 and GoAmerica.net

During 1997, I was one of the first people on Long Island using AT&T Wireless's CDPD service. I remember my efforts of sending emails while driving around New York City using a very non-intuitive user interface: the 12 buttons which where on my mobile telephone. While it was a very cool to have a telephone that was both ringable and pingable, for people who wanted to reach me via email, I still had to give out a separate email address. This was frustrating since I had no ability to check my corporate email while I was in transit.

As we migrate to an "always on" world, I'm sure there are other readers of The Pulver Report who can relate to the anxiety I start to feel when I'm away from email for lengths of time during the work day. The withdrawal effects for me are worse than any caffeine headache I've ever suffered. The lack of knowing "who is trying to reach me" creates a real feeling of being lost. (This is my own fault; I've stopped using voice-mail for myself in my office, and messages from people who have called me are emailed to me by my office.).

Fast forward a few years and for the first time in recent memory, I've stumbled upon a personal change agent for myself. With the advent of the GoAmerica.Net ( http://www.goamerica.net ) wireless ISP service and the availability of the RIM 950 ( http://www.rim.net ), New York is is starting to feel a little like Scandinavia. This is because of my availability to be an active participant in the world of the wireless internet.

Requiring only a single "AA" battery, GoAmerica.Net enables me to have direct access to send/receive email using my POP3 account and have the ability to be mobile not only around my office, but around the metro NYC area and most cities in the US I've visited as well.

In my office, I'm one of the only people who still use Telnet and PINE to read my email. In fact my email client PINE happens to be the client I use for my "instant" "messages". The good news is that while the rest of the world gets effected by the new strain of Microsoft email viruses, there haven't been enough users of PINE to attract the mainstream hackers, thereby leaving the PINE user community relatively safe. While I have no easy way to read attachments and have to take extra steps to read them, I'm used to it. Others less retro than myself may have a hard time not being able to view Excel spreadsheets or Word documents. The interface on the RIM 950 works for me because I've been living in a "text only" world.

For those of you who have seen but not used a RIM 950, the small keypad on the device reminded me of the original PET keyboards from the early 80's. Having used the device, my fingers fit fine on the tiny keyboard, and I find the user interface to be "usable". Reminding myself that this is a next generation end point and not a replacement for my portable computer, I can appreciate how powerful an affect this kind of appliance could have as time moves forward.

Now that GoAmerica.Net service is available, I'm waiting for the time when I can order Wireless ISP service for my car. This way, when I'm using a GPS system in my car, I can also share the 5 inch in-dash screen for my text based email. I just hope that owners of wireless automobile email systems choose to read and reply to their email only when their car is not in motion.

This said, I'm looking forward to the near future where family members and friends can export their personal voice fonts so that when a new message arrives, instead of me reading the message, the message is read to me in the voice of the person who sent it. This would allow me to be more interactive while spending time in the car and instead of listening to the news or music I can turn to a new station on my radio - my personal email station. For email replies, I'm pretty sure the right technology exists today for the speech-to-text translation and sending of the email reply while I'm driving.

Feedback from April 17th issue: ATM/MPLS clarification

In response to a portion of a story included in the April 17th issue of The Pulver Report, "QOS Happened", a reader who is a frequent speaker at the VON Conferences writes:

A minor nit:

> 3. QOS Happened

> > Overheard in the halls were the discussions QOS already
> happened in the VON world. ATM works. Quite a number of
> service providers have already deployed ATM.
> Over-provisioned ATM Networks seem to be the default Network
> Design of 2000. While some service providers are waiting for MPLS and
> DiffServ to be delivered over the next 9-18 months, those who didn't
> wait deployed ATM and are now servicing the needs of their customers.

The nit is "over-provisioned". In ATM networks, we have so called "hard" QoS. This means there is an actual reservation for resources, end to end. In fact, when a connection is established, we talk about a "contract" for bandwidth.

In ATM networks, in fact we sometimes see the deliberate overbooking of some resources when operators realize that not all customers use all of their "contracted" bandwidth. In general though, ATM networks are not over-provisioned. That term, also dubbed "over engineered" is usually applied to IP networks that promise voice services. The hope is that there is enough bandwidth in the network so that no significant congestion will break up voice traffic.

The problem, as many operators are discovering, is that it's close to impossible to actually over engineer networks in this way because of a problem called "focused overload", where traffic, instead of being distributed evenly across multiple paths is instead sent down an overloaded path because upstream routers don't know (yet) that it is overloaded.

Several operators who brought voice over IP systems into lab trials have postponed efforts to deploy large scale VoIP systems in favor of ATM systems because of this problem. Very large scale VoATM systems are being deployed now with guaranteed QoS. The amusing part of this is that ATM has always had an end-to-end bandwidth reservation system, the sort of system the IP gods have declared "not scalable". The traffic policing, queuing, scheduling and traffic engineering mechanisms in ATM have served quite well to manage these QoS contracts.

We are quite optimistic that the QoS mechanisms in MPLS will allow us to deploy large scale voice systems, even though these systems will be based on the relative priority "DiffServ" model. Such systems are not quite ready for prime time yet.

DSL needs a bit of PacketCable

After listening to DSL companies recently, I was surprised to discover how little adoption has occurred of VON technologies. This is largely due to a focus on supporting carriers with legacy central offices. I agree that it is a market segment, but the implications are that too much value has been placed on what exists. It has also done the disservice of focusing companies on GR303 over IP trunking as opposed to ISDN PRI or straight IP.

While I recognize the need to focus on companies with market share, few in our space are advocates, and the implication is not good for call management solutions. Many in the VON industry have adopted ISDN as a model for managing a network through a third parties network. As softswitches develop for ISUP models, this is an easy solution. Worst of all, the GR303 activity seems to be suffering from poor interoperability. So here we are in familiar territory a quasi - standard that is not providing the benefits of interoperability.

CableLabs' PacketCable had an impact on revising the DOCSIS specification. It seems that DSL forum could benefit form a group looking at VON technology. Another concern is that the market dynamics are changing. Recent FCC rulings has priced T1 cheaper than DSL circuits for CLECs. This makes the Integrated Access Device market hard to target. Recent conversations have indicated a CLEC market that has not caught the VON fever. We hope to see these folks come to understand that whether VoP is on ATM or IP, the distributed network looks the same, and the end user benefits from IP at the edge.

Trends in Instant Messaging: Feedback from our College Intern

College students are notorious early adopters of all sorts of Internet technology - from MP3 trading to ecommerce, thanks to ubiquitous always-on high-speed access. To no surprise, my intern reports that Yale students are firmly enmeshed in the culture of instant messaging. The technology, especially as found in AOL's Instant Messenger and ICQ offerings, is used not just to keep in touch with high school friends spread across the country, but friends in the next room. The benefits are numerous, the service is unintrusive and even serves as a 'tickler' about people students need to talk to. More importantly, the away message functionality provides a passive means of announcing location and availability. It's a prime example of why instant messaging will inevitably be entangled with VoIP in a converged IP future.

Presently, students set away messages even when they are at their computer to indicating what exactly they're up to and to give a sense of whether it's appropriate, based on the reader's sense of urgency, to drop by, to call, or to message someone. However, the lack of integration is painfully obvious. Students act as their own gatekeepers, sending their phone number to friends when they're available and want to be called, or announcing when they themselves are on the phone so that people don't try to call or IM.

The implications for the wireless future are significant. There must be some way to transfer this widely-popular away message functionality, particularly the ease with which a text message is entered into the client software so that people attempting to contact one another can get a real sense of how available a person is and by what means. It could even be converted via text-to-speech to form a sort of dynamic answering machine, maybe even using technology to mimic the person's voice. This idea of exporting state and profiles, perhaps even dynamically based on cellular location sensing, will be increasingly important as varieties of access methods proliferate. Alternatively, when a person has announced that they cannot be bothered by a voice conversation, a voice call can be converted to a text instant message.

Voice over IP has not quite caught on in the college market. Students seem unwilling to sacrifice quality and convenience just to converse for free over the open Internet. There has been some success from providers such as I-link ( http://www.i-link.net ), which provide a somewhat more customary degree of ease. Yale's Director of Network Operations foresees implementing SIP on the campus network within two years, and the school's Director of Information Technology Services is looking closely at deploying a private wireless cloud to facilitate both data and voice. Already, many students have turned to private cell phones for there convenience, privacy and cheaper rates. The viral success of instant messaging and cellular phones, in contrast to that of consumer VoIP plays, should be a lesson about the importance of the application layer and the direction it might take.

The Pluses and Minuses of Research and Development.

One industry trend that needs to be recognized is the exodus of R&D talent in various carrier companies to new startups or specific suppliers. In effect, what is going on is that the Industry is admitting that R&D is something they do not do well. Not because of the talent but do to the lack of commitment. It is sort of a reverse of "Not Invented Here" [NIH].

Carl has discussed from seeing Lab demos at various carrier R&D facilities that many are "cool solutions". But, the carrier often couldn't get the coolness out of the lab in a way that made it operational. Countless times I have watched engineers smile gleefully at simple call completion during demos, but what is needed is an easy to adapt platform for a number of services. The carriers are keeping tactical people who can aid operations in providing service while shedding people with vision.

The assumption is they will solve the issue of distinction by marketing...either bundling many services together or by selecting a key feature to implement to gain "stickiness" in their application.

I am not sure what the ultimate result is, but I am impressed with the VON companies that are taking advantage of this phenomena and hiring the available talent.