How To Build The Ultimate Network

What organizations would want their networks to be, if only they had all the money, time and expertise in the world, is hardly a mystery. Indeed, in a way, the ultimate network is really about nothing more than the Olympics' motto "citius, altius, fortius" rephrased as "faster, more efficient, more reliable." Just how you go about building this network, however, is another thing entirely.

"It exists in utopia," says Info-Tech Research analyst Carmi Levy. "In reality, there's no such things as the 'ultimate' anything. The only way to achieve it is in the lab, and even then, that's probably not even realistic."

Although the ultimate network exists only in theory, what is realistic is to make it a target, Levy says. The best thing any organization can do is to take a tip from Friedrich Nietzsche's superman, whose "reach forever exceeds his grasp."

That's good advice, perhaps, but it begs the question of how you actually go about planning for the ultimate network, even if it's a goal you can approach without ever actually achieving it. Is it a question of spending bundles of money -- just like in the days before the dot-com bubble burst -- on the hottest equipment, infrastructure and software?

Well, top-end technology never works but Levy says that focusing on technology obscures the real paths to the ultimate network. "It really isn't a technology issue," he says. "These things never are. We're always throwing things onto the network without thinking about how they affect the network."The key is to make sure that the architecture and the network roadmap are planned from the ground up. "Instead of a hodge-podge of processes, things have to be there for a reason," Levy says. "If they aren't, then they have no business being connected to everything else. You don't build a building without blueprints, and you shouldn't build a network without a roadmap."The problem, of course, is that organizations seeking the ultimate network are probably already dependent on a network that is decidedly un-ultimate. Building the best is a whole lot easier when you start from scratch, and it's easier to build a house from blueprints when there is nothing already there, but only start-ups have the benefit of a technology greenfield. On the other hand, Levy says, you can only pin your problems on legacy technology for so long.

"At some point, you have to stop blaming legacy technology," he says. "You have to integrate the good and the bad in you plans and move forward. No one can divine the future, but you can prepare for it by planning from where you are and building with open standards. If you have a clean, modular network, even if it's old, you can upgrade."

Standards, in fact, are a key component of any network that claims the title "ultimate." With the growth in server virtualization and storage networking, for example, many organizations have found themselves running up against compatibility and performance issues that can bring core processes to their knees. EMC director of technology and analysis Ken Steinhardt says "The new thing, and 'first to market' aren't necessarily good things if the standards aren't finalized. Standards that provide interoperability are key."

With that all in mind, the first step is to assess what you've got in terms of hardware, cabling -- even the physical plant where your network assets are installed. "If nothing else, you should do an inventory," Levy says. "Then you can identify your points of pain -- where are you constrained, performance-wise, and how can you change that?"

Ultimately, the quest for the ultimate network should lead up to a short list of business impacts fro insufficiently or inefficiently deployed technologies and processes. Make no mistake, a bigger network pipe and faster hardware could well be in your future but, Steinhardt says, "the key is to have the full chain; you're never going to be better than the weakest link in the chain."Indeed, the mistake that most organizations make, according to Levy, is that they go shopping too soon. "All the technology in the world isn't going to help you -- and it'll probably hurt you by wasting money -- if you don't know what you need and how you're going to deploy it," Levy says. "Again with the architectural metaphor, you don't build your house if you don't know what kind of land you're on. You need to know your topology and re-arrange it if necessary; if you need to segment then segment properly."

Building the ultimate network, then, is about planning the ultimate network. It is, above all, about doing your homework, knowing what you have, what you need, and what you can afford to spend. "Then you go shopping," Levy says. "Bring in the vendors that you need to get things done, and show them the plan."