Tuesday, October 22, 2002

Speaking of Changes . . . I spent some time with the folks at Flamenco Networks recently, and the visit really got me thinking about the transition networks will undergo over the coming years. That transition is only just beginning, and it’s going to cause changes not just in infrastructure technology, but in how enterprises buy and manage that technology. Flamenco is one of many vendors -- including Confluent, Grand Central, Talking Blocks, Vordel, and a bunch of others -- that are focusing Web services management. On one hand, the focus on managing the problems Web services will inevitably create makes perfect sense. On the other hand, however, at least some of these products blur lines between the traditional “application architects” and “network and telecom architects.” And that creates transitional issues for vendor and customer alike. For example, Flamenco's products help enterprises set up, manage, and tear down "connections" between Web services end points. Operating at a higher layer than basic network services such as VPNs, Flamenco's products add a new layer of “application network” functionality. They authenticate end points and provide guaranteed delivery, monitoring, auditing, and other services that create a more secure and reliable network. Flamenco uses a basic proxy architecture. And one of the primary things I like about their approach is that they help end points bootstrap peer-to-peer communications, rather than force all communications to pass over a third-party network. But who's the primary buyer of such products? Today, enterprise network architects are getting pushed more and more into the security and management arenas. They're having to manage VPNs. They're managing firewalls and other security services. And they're moving up the stack into application issues with deployments such as Voice over IP (VoIP), storage area networks, and caching and load balancing for Web Servers. But are they ready to buy and implement a new layer of application network functionality? Do they even understand the impact Web services will have on their networks? Traditionally, enterprise application architects have dealt with application integration issues. And network architects often gripe about their tendency to bring up new applications without telling anyone about them in advance, thus creating problems for the network management team. What happens to that dynamic when Web services create integration issues that span internal and external boundaries? How can enterprise application architects start working with network architects in a way that anticipates these problems? Perhaps Flamencos deal with British Telecom is a leading indicator, at least in terms of products. BT licensed Flamenco’s technology as the basis for a new class of networking services, intending to position the Flamenco-based “application network” functionality as a higher layer service that compliments its existing connectivity services. Carriers looking for life beyond bit hauling should certainly see opportunity in helping enterprises create reliable application networks, and it won’t be surprising when other carriers follow BT’s lead. But how quickly will that happen? And how big is the market for BT’s services today? Simply put, we’re in the middle of the transition, one of those liminal moments in which the old way is dying but the new way has yet to emerge. That means both vendors and enterprise customers will have to deal with cross-organizational selling and implementation exercises that would have been unthinkable just a few years ago. Enterprise application architects and network architects need to recognize that trend. The companies that encourage higher degrees of cross-pollination across these traditionally separate responsibilities are likely to have an easier time managing the transition. 5:54:35 PM

How Do You Rip Off a BandAid: Slowly, One Hair At Time, Or In One Fast Motion? It's been amazing to watch the telecommunications market decompose over the last year. The letter posted here makes the case that, since the old has to die before the new can emerge, we should hurry up about it. (Thanks to Dan Gillmor's blog for the link.) The letter urges the FCC not to prevent the inevitable, and thus prolong the agony, but to let the dying carriers die, and die quickly. While folks here like David Passmore and Ray Keneipp follow the telecommunications market more closely than I, this point of view certainly makes sense to me, at least from one perspective. On one hand, this is a case where the Darwinian characteristics of the market should be left to their own devices, not artificially engineered by the government. On the other hand, however, the cynic in me wonders how the government can withstand the pressures of a big-money lobbying effort and an uneducated public. And, as Ray Keneipp points out, such arguments "ignore the political realties of having an ILEC fail, leaving thousands of people without phone service." One thing's for sure -- it will be interesting. 4:24:40 PM