Does thinking about a networking lifecycle make sense?

Does thinking about a networking lifecycle make sense?

Summary: The industry is used to considering system, storage, and application lifecycles. Companies routinely cycle through new requirements, standing up something new to address the requirements, running for a time, and then retiring equipment or applications. Does it make sense to do the same with networks and networking equipment?

TOPICS: Networking

I enjoyed a delightful conversation with Stephen Garrison of Pica8 about virtualization in general, network virtualization and the trend towards software-defined networking (SDN). We touched on so many interesting topics that I couldn't possibly get everything down in my notes. So, I'll focus this post on one very small aspect of the overall conversation — networking lifecycle.

The system and storage lifecycle

Garrison wrote a blog post, Does Network Lifecycle Management Make Sense?, that considers the issue. His premise is that IT architects routinely examine the performance, cost, power consumption, and heat production of their systems to learn if newer devices would be more effective. The same sort of examination is undertaken for the company's storage resources.

Some companies have decided to constantly replace equipment that is no longer cost effective with something new, even though that equipment hasn't been in place the typical two to three years the industry considers standard. Instead, they operate in a constant state of equipment lifecycle management to optimize their return on investment. The same, Garrison pointed out, is not true with a company's networking equipment.

Networking equipment has always been different

Networking equipment, on the other hand, often stays at the top of the racks for quite a bit longer, sometimes up to five to seven years, even if something else is available that offers much better performance and reliability and would reduce the company's cost of operation.

When Garrison asks clients and potential clients why, they often point out that the networking equipment is very difficult to stand up and operate. The equipment often has been designed to use a specific, propriety command-line interface. Companies find it too complex or costly to do much more than install the equipment and use it until it breaks, regardless of whether something better is out there.

Wouldn't it be better to think in terms of a network lifecycle?

Garrison proposes that it would be wise for companies to rethink their network infrastructure and move in the direction of networking equipment that is easier to manage, update, and operate. He suggests that the newest generation of networking software and network switches make it possible to use technology such as the Open Network Foundation's OpenFlow. A virtualized network, a network that separates command and data to create a much more manageable and flexible environment, would make it possible for companies to do with their networking infrastructure what they're already doing with their systems and storage.

If one considers the cost of a company's network infrastructure as part of the overall cost of IT, it would be wise for companies to start evaluating their network infrastructure using the same criteria used for their systems and storage infrastructure. Moving in the direction of a virtualized network would serve to make the implementation of a network lifecycle more easily possible.

Topic: Networking


Daniel Kusnetzky, a reformed software engineer and product manager, founded Kusnetzky Group LLC in 2006. He's literally written the book on virtualization and often comments on cloud computing, mobility and systems software. In his spare time, he's also the managing partner of Lux Sonus LLC, an investment firm.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


1 comment
Log in or register to join the discussion
  • It's more than switching costs

    The biggest reason people leave networking gear in so long is that it is prohibitively expensive to buy. The pricing structure of the networking industry at large has been skewed because there is a virtual monopoly. With little competition, Cisco has been able to keep demand high (and hence prices and margins).

    SDN is interesting, but not just because of the technology. It is leveling the architectural playing field, which lowers the barriers to entry for new players. This creates more competition, which will exert downward pressure on pricing and margins.

    If devices are cheaper, they don't have to stay in place quite so long to justify the expense. Sure, the switching costs matter too, but even if replacing a box was easy, it wouldn't happen previously because it was too expensive to rip out.

    The biggest thing to change going forward is the organizational inertia around buying patterns. People need to get emotionally and intellectually prepared to handle a different way of buying and refreshing gear.

    Mike Bushong (@mbushong)
    Mike Bushong