Convergence, Chaos, and the Laws of Network Entropy

By ​Charles T. Cranford

In our first two articles we covered different approaches to communications that you’ll encounter based on industry backgrounds [Polyglots, Plutonians, and Mercurians], and how to make sure you don’t fall for the one-size-technology-fits-all [Pac Man, The Sirens Song of Simplicity, and Integrators/Consultants]. This time we look at the future of Nirvana [still without Kurt or Cortney], named Convergence – one network will do it all!

This utopian goal has been espoused by more and more vendors since the mid 1990s when a small software company launched a low-cost Voice-over-Internet-Protocol (VoIP) application. It hasn’t been the same since! Company after company began to launch their “projector-ware” architectures-of-the-future product presentations. Trade show floors were filled with Booth “Babes and Bulks” singing out (or passing out) information on how these networks of the future would combine voice, data, video, and facsimile transmissions – all with their own speed and response requirements being fully met. Vulture, uh, “venture” capitalists were seen throwing money at companies whose business plans were written on the back-sides of napkins at high-end bars and restaurants. But the best was yet to come…reality!

While it is true that you can put voice, data, video (and almost anything else you’d like) onto an IP-based network, it is also true that you need to understand there are user expectations to meet. And expectations will separate the successful applications from the lab-experiments-gone-bad. In this case, it is important for both vendors and users to both have technical/protocol-knowledge – kinda’ like the saying “being forewarned is being forearmed.” Or wearing “belts and braces” (as the British like to call the U.S. equivalent of “belts and suspenders”…strange “English” terms…) which means having yourself doubly-protected.

Voice and video applications, of course, are time-sensitive applications and users expect to have these packets delivered with first-class service. Email, web-browsing, and file-transfer applications can suffer back-of-the-plane, coach-class service, since the user won’t notice the same level of delay that a voice or data user would (“Look, Homer, that email was 750 milliseconds late getting into your Inbox!” probably isn’t going to be heard around the office).

By this time astute users of IP-based services are about to turn the page and find something that is “new” to them…because the above issues are almost tautologies to them. But while people know that IP can “do all”, what most folks forget is that you need to know more than IP – the Law of Network Entropy is at work even in your networks.

The Law of Network Entropy (or measure of “disorder” in a system) simply states that a network administrator will need to know more than “just” IP. You will need to know which Layer 2 you will use (worldwide, Ethernet is the clear winner), however, your LAN’s Layer 1 alternatives (fiber, coax, twisted-pair wire, or wireless) continue to grow in complexity-and-disorder…uh, I mean “flexibility”.

Flexibility/Complexity/Disorder is easy to see just in the alternatives for Fiber/Fibre (with nods to the U.K….), for example. Fiber/Fibre can be implemented with both short-range and long-range services as well as single-mode (which, paradoxically, sounds “limiting”) and multi-mode (which, also paradoxically, sounds “more-better”). And this is before you decide whether to have single-channel service, multiple-channel service (CWDM, or Course Wavelength Division Multiplexing) or “lots more channels” service (DWDM, or Dense Wavelength Division Multiplexing). And that’s just your fiber options…don’t get me started on the multitudinous ways that “wireless” services can boggle the mind!!! The numbers of choices can be “Inconceivable”!

On the WAN-side, your service-provider’s services vary from Private Lines to “virtual” private-line and “virtual” private-network services (hint: remember that anything following the word “virtual” means “it’s a LIE”…it’s not really a private line or a private network!!!). Frame Relay (FR), Asynchronous Transfer Mode (a.k.a.: the ATM that vendors can’t get money out-of), Point-to-Point Protocol (PPP) and Multi-Protocol Label Services (MPLS) will be the likely alternatives among your WAN-provider’s bag-of-uh, uh, Layer 2 “services”.

Also offered in many advanced networks is the Layer 3, IP-based Virtual Private Network or Software Defined Networks (now we’re back to IP-does-everything), but, again, the Layer 3 service requires the right Layer 2 (see above paragraph) and Layer 1 access and backbone selections…

So what does this mean for today’s technical savants-to-be? It means that you have to understand that:
1) Network Services will focus on Layer 3, IP-based networks (given!)
2) Networks need more than Layer 3 (hence the names Layer 2 and Layer 1 that live “below” them
3) Network users (us!) need to know and understand what these alternatives are for our WANs as they provide the “underpinnings” for the Service Level Agreements (SLAs)
4) The Law of Network Entropy continues to offer us more confusing, uh “flexible” alternatives from when we may choose
5) The “One Protocol Does All”-dream is just that…a dream. So keep learning more Protocol “Languages”, so you don’t get caught like a tourist dropped-off in a foreign non-American-speaking country (like the United Kingdom!)…you need to know ‘what’ means ‘what’!