""Howard C. Berkowitz"" wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... > Perhaps your most important point is at the end, but I'd like to > amplify on it. The idea of a converged public Internet is probably > not achievable. I prefer to call service providers that move packets > "IP Service Providers," as opposed to "Internet Service Providers." > It's more accurate, and reflects the very different availability and > QoS requirements of applications, balanced against costs.
Yes, absolutely. And it also has to do with security as well - especially as it relates to accountability. On the Internet, people have the presupposition of anonymity which, while important, also can be used a shield by criminals like hackers. A private IP network carries little presupposition of anonymity, so if you're a customer of a private network and you're committing mischief, it is much easier to find out who you are and terminate your connection. There is no God-given right to a private IP network connection the way there is with the Internet. > > Right now, a lot of world-class router designers are unemployed or > underemployed, because there is so much optical overcapacity that > sophistication isn't needed, especially with private networks. This is also true and hurts not just world-class designers, but network engineers of all stripes. Moore's Law churns inexorably and so do advances in optical technology. Things like QoS matter less when you can cheaply throw bandwidth at a problem. Things like voice over IP matter less when it's inexpensive to have lots of separate networks. The ability to carefully engineer and tweak your network is relatively less important when bandwidth is plentiful and it therefore doesn't really matter if your traffic takes suboptimal paths. > Much > of this, of course, is the current economy, which I do expect to turn > around. From my IETF/IRTF work, I do know that the current global > routing system isn't going to grow forever with the BGP paradigm, and > the best replacement is still a research problem. Luckily, I'm able > to keep a hand in that. I think more thought needs to go to how to turn an IP network into a profitable service. For all the problems of ATM, one indisputable thing about it is that it actually creates profitable services. I think there has been too much emphasis on developing 'cool' IP technologies and not enough has been made on creating profitable IP technologies. > > We are a long way from having every application run on a commoditized > transport. I'll freely say that more of my income, these days, comes > from both network and application architecture for bleeding-edge (a > phrase the surgeons HATE) medical systems. Now, some people here say > you need host as well as network experience. While I'm reasonable at > UNIX, there's also the aspect of being able to communicate with the > users of particularly challenging applications. I speak fluent > Doctor, which helps greatly, and can actually contribute to the > clinical application designs. > > Don't assume that you necessarily have to have extra computer skills > (e.g., server administration). Understanding an application area > from its user perspective can generate lots of work, be that > application telephony, medicine, law, etc. I have a friend who has > developed a specialty in automating car dealerships, and he has more > work than he can handle. My point was not to say that you necessarily need to have strong server knowledge. My point was that you needed to stretch beyond R/S. Anybody who tries to live on R/S alone is living on borrowed time. Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=59347&t=59347 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
