The three real issues in networking’s future

Beyond the hype of white papers and other vendor-driven material, IT, network and telecom managers struggle to define the real issues of networking. In this excerpt from his recent BCR article, Tom Nolle, president of consulting company CIMI Corp., points us in the right direction.One of my clients made what I think was a key comment in a recent meeting. "The problem we have in the industry today is that we're focusing on the issues that advertisers want us to focus on and not the issues that are really important."I hate to admit it, but the client has a point. You can read "networking's future" white papers galore, but when you examine the contents, they're really about how a given vendor's solution can be applied to some problem clearly invented by the same vendor. This begs a simple question: What are the real issues in networking today?Issue 1: SettlementIn the past, network services were provided by players that were all offering pretty much the same thing but in different geographies. This created a simple proposition for co-operation: when you connected a service over somebody's infrastructure you paid them a fee for use. But the Internet created a new vertical stakeholder stack in communication services, which also created a new set of problems. The debate we call "net neutrality" is the manifestation of the settlement change. If the consumer pays for Internet access and the content providers pay for Internet access, why can't everybody just get as much as their access conduits will support? But there's a big difference between building a network to support consumer Web access versus one that delivers HDTV. There is no question that infrastructure costs will rise as we increase our appetite for downloaded material. Net neutrality says Internet-delivered content won't provide additional settlement from the content side to the access side. That either forces consumers to pay the additional cost of high-QoS access, or it forces content off the Internet and onto IPTV parallel networks. Either approach is workable; both have consequences. The debate should be over whether these consequences are worse than those created by content settlement on the Internet.Issue 2: Operations Cost For network operators, the IP-convergence paradigm was based on the simple notion that if multiple networks were eliminated in favour of a single IP network that supported all services, total network operator capex would be lower. While that's likely, provider experience so far has been that opex is higher with converged networks. One mature converged carrier told me that their VoIP complaint rates were more than five times PSTN rates, and that their VPN complaint rates were 3.5 times that of legacy private network services.A related problem is that of "touch," the ability of a provider to offer a business customer a premium support experience. Many providers have commented that IP networks, with their adaptive behaviour, make it difficult to write a traditional service-level agreement. The consequences of that are dire, according to one of the providers I talked with: "If I can't replicate my frame relay SLA with an IP service, I have to discount the IP service to make up for the loss. That undermines the whole economic justification of offering the IP service in the first place." Clearly, offering a comparable SLA is technically possible, but the operations processes are too costly or not available. Issue 3: Federation We all know the network of the future will host software, storage, content, and so forth, just as today it hosts simple connectivity. Today, we create "federations" of connectivity: private networks, virtual or otherwise. In the future, we'll create federations of applications and service features hosted on servers instead of embedded in network gear. How will we do this? A simple example is application connectivity. You have a private network today, and if your provider is hosting one of your applications in the future, that application has to be a "member" of your private network. That alone creates some significant security issues, as well as questions of how that hosted application connects with other applications that you retain internally. If you host a second application with someone else, and that second provider becomes part of your private network, are the two providers now visible to each other? If so, have you created a hole where other people can route traffic through your network? If not, how do those applications share data? Service features are similar problems. You have VoIP and so does your caller. Each of you has voice mail, caller ID, call forwarding. All of these features are hosted by your respective providers. Do they co-ordinate correctly?The simple answer is standards. But the number of standards groups grows every year, and every year most of the groups are taking on new topics. Many overlap each other. The standards don't always perform as promised. We live in a world dominated by commercialism, and the stuff that gets attention is the stuff that affects commercial interests. More than that, it's the stuff that has a positive commercial spin, meaning that somebody can make money from it. Issues that, if they are not effectively addressed, reduce the total amount of money that changes hands, are not visible because nobody benefits from exposing them. This article is reprinted by permission of Business Communications Review (http://www.bcr.com/).

The three real issues in networking’s future

News | 05/16/2007 1:02 pm EDT

Unlock all the Canadian telecom, broadcasting and digital media news you need.

FREE TRIAL

SUBSCRIBE

Reuse & Permissions

Inside The Wire Report

Information

Hill Times Publishing

Get in Touch With Us