Reliable VoIP

Since communities first established storehouses of grain to provision against future famine, people have organized government to prevent shortages of life’s essentials.

Electronic communication, in all its forms, has become essential to our continued prosperity, liberty and social advancement.

Fundamentally this society is in transition from a robust redundant and managed telephone system optimized for universal voice service to a faster and more diversified collection of unmanaged communications designed for any kind of data.

While the benefits of this transition are numerous, the trade-offs have received less attention.  The key issues in reliability flow from two fundamentals:

First, the physical architecture of much of the internet is optimized for cost and not reliability. For technical reasons given the art of the time, the original phone system was deployed in a highly parallel manner with separated wire pairs for each line running back to a local central office. Even in an extreme disaster, such as a tornado, service was often available or quickly restored across a wide area. This is no longer true in all modern deployments of internet and VoIP today.

Second, the software and protocol architecture of the internet favors in-band signaling, i.e. combining data and signal (control) together. So for example if you compare and contrast the history of the H.323 protocol (having its roots in ISDN) with the more recent SIP protocol, there is evident a modern architectural movement toward greater convergence of data and control.

Technical choices are being made in favor of convergence, cost and features.

To a modern designer avoiding convergence violates the ideal view of all bits as equal as converging data is highly attractive if you assume reliable delivery.

The social issue of who is responsible for assuring reliability is not captured in today’s economy. The complexity and costs of high reliability are disfavored for events beyond the ordinary recapture of revenue.

At issue is the social deferral of the costs of emergency. The commercial market with current policy tends to disfavor adding costs which evade recovery.

The issues above are not confined to wired telephony. They extend to wireless as well.

When answering the question─ how are radio towers provisioned? ─consider whether the answer takes the data on a path through a vulnerability. When it does, the tower is no longer an independent reliable backup.

Thus, it may be prudent to ask and consider the following questions:

  • Is there a consensus of knowledge about the physical reliability of the internet in handling emergencies?

  • If not, what projects might be proposed to bring the value chain to a common point of understanding?

  • Is there a consensus of knowledge about the actual redundancy of converged communications?

  • If not, what projects might be proposed to create a common view?

  • Is there an agreed sense among all constituencies on the best practices for overflow and capacity planning of the internet and VOIP?

  • What might be done to encourage industry and the public to prepare for communication in the event people are stranded and unable to get to their customary and approved means of communication?

From A Statement Offered In Support Of Testimony in Washington, D.C.
In The Matter of Planning For Social/Governmental Emergencies

Jonathan Zar is Secretary & Outreach Chair for VOIPSA, the VoIP Security Alliance. VOIPSA represents 100 organizations and over three thousand of the world’s experts in converged media security. Mr. Zar would like to acknowledge the valuable contributions informing his statement from Mr. Robert Simkavitz and Mr. Philip Walenta of VOIPSA. Mr. Zar’s words are his own and he has offered his statement as a private citizen and not in his official capacity as a spokesman for VOIPSA.

Copyright (c) 2006 All Rights Reserved
Permission Granted To Reproduce Intact Citing This Posting