December 1999

Just The Fax
I checked out the jfax site mentioned in the October '99 Reality Check in Internet Telephony� magazine and it looks like it is no longer free. There is a $15 setup fee and a monthly fee of $12.95 (the monthly may be off a little. I'm going by memory).

Am I missing something?

—Greg Starr,
Information Systems and Technology Manager
Milwaukee Symphony Orchestra

Robert Hashemian responds:
No, you are not missing anything. By the time my column hit the street, jfax.com’s service became a paid one. I still think jfax.com offers a good service although I liked it much better when it was free. My apologies for any confusion. But wait! After a bit more digging (in late October), I discovered that jfax was still offering a free service called JFAX.COM Free Fax Plus. Check it out for yourselves!

Having worked for a unified platform provider, Comverse Technology, I can certainly understand your position. However, consider this from the standpoint of a telecom carrier. The paperwork we require signatures for is as follows:

• Purchase Order (PO);
• Design Layout (DLR);
• Letter of Authorization (LOA);
• Contract for Carriers;
• Sales Orders;
• Termination/Disconnects;

…and the list goes on. Just like the mortgage business, almost every transaction involves the binding legal commitment of an authorized signature and herein lies the problem.

It appears that I now have more paper then ever! Oh, well your article provoked thought.
– Steve Lynch

In The Name Of Balance
In reference to Ken Dumont’s article, Balancing Voice And Quality On The LAN (Internet Telephony� October 1999).

The only thing I could think of when I read this article was “wishful thinking!” That is, it is “wishful thinking” on the order of magnitude of IBM mainframe guys trying to dismiss the advent of the PC/LAN world back in the early ’80s. Although there is nearly every line in his article that I could take issue with, in the name of brevity, I will only hit on a few of his more obvious mistakes:

PC Reliability & Power Loss. There is no reason why PCs cannot be designed (and are) to be as fully redundant as a proprietary PBX. Also, a PC-based platform can be made more reliable by having only one mission-critical application running on it rather than several at a time. With the advent of meshed fully redundant design for IP voice networks, the network itself can be made as reliable as the proprietary stand-alone PBX. It is true that power backup may be required in the network. That can be accomplished locally and from the switch itself. Although power issues traditionally have not been a concern for many data networks, this will be solved for the new integrated voice, video, and data architecture.

Voice Quality. Delay, jitter, and echo are being solved with today’s IP phones utilizing today’s new DSPs and hardware designs. Superior voice quality on IP networks is demonstrable today! In fact, the “old world” voice network is limited to the equivalent of a G.711 or 64k call. With enough bandwidth, there is no reason that the voice quality can not exceed 64k by utilizing, for example, G.722 (64k–256k)!

Legacy Migration. Vendors today are selling IP telephony products compatible with existing PBXs. For example, Cisco’s AVVID will interface with virtually all PBXs and interoperate completely with Octel’s (a division of Lucent Technologies) voice mail system. The design goal is to eventually interoperate with most third-party VM systems. In addition, the advantage of this open architecture that utilizes SMDI and open standards is that many new third-party apps will work on this system that cannot work on a proprietary PBX. Who has the best long-term investment protection?

Investment Protection. It is not necessarily true or even apparent that most customers have staff on hand that know how to program their PBX. Most PBX vendors make a great deal of money supporting and programming their “proprietary” boxes. In fact, the open standards world allows companies to utilize their existing IT staff to do most of the support of managing standard hardware (standard PCs and components), software (NT, Unix, and third-party apps), and network (IP infrastructure). An open non-proprietary network costs much less to support. Also, the price of proprietary PBX phones and cost per port has (as expected in a closed proprietary system) essentially stayed the same ($500+) over the last five years. Over the last five years, the cost of a switched port on Ethernet has dropped from around $1,000 per port (10/100 switched Ethernet) to about $100 per port in the last five years! One can expect the cost of IP phones to do the same! Finally, technology exists today that will allow customers to gradually migrate their large PBX networks to IP. For example, a gateway will allow the existing PBX phone to emulate an IP phone. This way, much of the hardware in the PBX network can be gradually transitioned to IP without “forklift” upgrades.

Life Cycle. It is not strength but a weakness that the PBX has a five- to 10-year life expectancy. That means that a proprietary closed system will not essentially change for that time frame and the vendor of this equipment will be able to charge much higher prices. In the New World of Moore’s Law, the 18-month life cycle means that customers will see innovation in hardware and software on open platforms and all of this lowers the price (as exemplified by the Ethernet switching example).

It is true that there is much to be done before we look at the IP infrastructure as the functional equivalent of the PBX. The good news is that in the very near future, the “new world” infrastructure will not only equal, but exceed the traditional PBX in reliability, quality, cost of ownership, and a rapid evolution of applications that will bring us to the point where we can finally see what those visionaries were talking about when they coined the term CTI.
– Alan Liss, Cisco Systems