In search of supreme network performance? Try deploying VoIP. Considering its reputation for creating serious network problems, it may not seem obvious that VoIP can actually show you how to make your network perform better. But it can and here's why:
Your network is a relatively complex form of ecology. There are devices of all sorts like routers, switches, client stations, servers and the like. Then there are the interconnecting media and interfaces such as cables, NICs and drivers. And there is the traffic and its various behaviors such as congestion and burstiness. Finally, there are network configurations, from interface settings to routing, that define the individual and aggregate behaviors.
Network performance is a function of all these elements and relies on very subtle interplay between them. For example, very low packet loss (< 0.1%) on a very high latency link can result in performance degradation for applications using standard TCP -- an unintentional consequence of TCP's slow start mechanism over long network paths. No single element of the network is entirely responsible -- the low packet loss would be considered reasonable under any other circumstances.
Even the devastating problem of duplex mismatch is defined by the not-so-subtle interaction between two interfaces — neither interface is actually malfunctioning (although it might be argued that auto-negotiation has a fault).
So if your network is a kind of ecology, then, Voice-over-IP plays the role of the amphibian. It is well known that frogs are highly sensitive to toxins in their environment and their disappearance can signal a serious ecological issue. If you can keep the frogs alive and well, chances are that most everything else will be healthy and happy, too.
The same theory rings true for VoIP -- to keep that VoIP frog alive, your network has to be clean. Very clean. And coincidentally, that is the key to high performance of all your applications.
A majority of performance problems are rooted in the degradation issues that also impact VoIP. Networks that cannot perform at 100% keep applications from operating at 100%. There is no way to tune an application to overcome a bad network.
Typical sources of network degradation include:
- Lossy media (wireless)
- Bad cables or poorly seated cards
- EM interference and poor shielding on copper lines
- Duplex mismatch
- Bad NICs/drivers
- MTU mismatch (RFC 1911 non-compliant)
- Mix of duplex domains
- Unplanned bottlenecks
While most typical data applications will work reasonably well in the presence of most of these -- this is where the frog comes in -- VoIP will not. And end users will clearly hear these problems. Even small amounts of loss and jitter are noticeable, forcing engineers to either resolve the degradation issues or compensate with QoS mechanisms.
So, if you were wondering how VoIP was going to fix everything -- it won't.
The good news is that the pressure to move to VoIP will force the development of tools and technologies capable of resolving hard-to-find network degradations. Allied Business Intelligence estimates that enterprises migrating to voice from traditional circuits will create a $16.5 billion dollar IP-PBX market worldwide by 2006.
The bad news is the hard work has yet to be done -- only 12% of corporations have deployed VoIP and another 29% are expected to join in next year (Harris Interactive) -- and we don't have all the tools we need yet. However once we deal with the uniquely sensitive nature of VoIP, the worst will be behind us and we will have reaped the benefits for many other network applications.
So clean up your networks and keep the frog alive.
Chief Scientist for Apparent Networks, Loki Jorgenson, PhD, has been active in computation, physics and mathematics, scientific visualization, and simulation for over 18 years. Trained in computational physics at Queen's and McGill universities, he has published in areas as diverse as philosophy, graphics, educational technologies, statistical mechanics, logic and number theory. Also, he acts as Adjunct Professor of Mathematics at Simon Fraser University where he co-founded the Center for Experimental and Constructive Mathematics (CECM). He has headed research in numerous academic projects from high-performance computing to digital publishing, working closely with private sector partners and government. At Apparent Networks Inc., Jorgenson leads network research in high performance, wireless, VoIP and other application performance, typically through practical collaboration with academic organizations and other thought leaders such as BCnet, Texas A&M, CANARIE, and Internet2. www.apparentnetworks.com