I've spent so many months over the past few years examining the many dimensions and possibilities inherent in "as...
a service" that perhaps it's time to start talking about everything as a service.
I've always advised strategic assets should be owned, while everything else should be rented -- in other words, bought as a service. Many applications and certainly the data within an organization are strategic, but what about the rest of IT? Are we at a point where, for example, networks are really commodities, and because of that shift, purchasing networks and much of the rest of IT actually makes sense?
This is a question of vital importance to any organization today. IT dollars are always scarce, so why spend money that would be better deployed elsewhere? If we really can purchase networking as a service, it seems we should.
The network-as-a-service market and what it means today
Let's define network as a service for our purposes here. Purchasing a capability or function as a service converts what would normally be a capital expense -- with its associated, required budgeting and depreciation calculations -- into an operating expense. It's the classic buy-versus-rent analysis.
We might think of the network-as-a-service market as another form of virtualization, wherein equipment and its associated operating costs become the supplier's problem. The vendor does the buying, installing, maintaining and operating, and the customer gets the benefits that accrue from those operations. As a result, net costs for the organization, as well as the supplier, are often lower.
Let's add a few more benefits shaping the network-as-a-service market approach: These include flexibility; simplified growth and stability; and last, but not least, enhanced fault tolerance and resilience.
Network as a service is just like many other managed services that organizations are already purchasing, from distribution partners to building maintenance. The most important elements in making the network-as-a-service market decision, then, aren't related to networking technology or products. Rather, it's understanding operational requirements and how these evolve over time, as well as assembling a list of relevant best practices.
Complicating the decision is the ever-evolving nature of networking today. Functionality varies widely among vendors, and innovation shows no sign of abating. That's why it's vital for end-user organizations to carefully specify their functional and operational requirements, and not specific product features, when going out to bid with network-as-a-service providers.
It's also important that agreements permit regular operational reviews and allow the addition or subtraction of equipment and software as requirements evolve. Successful network-as-a-service market suppliers understand customer requirements are never static for more than a few months, and they will build ongoing adaptation into their business plans.
The evolution of enterprise networking continues
It seems very likely we'll see today's enterprise network become a small set of components: access points; switches that provide interconnect and power; and an enhanced firewall, which is what we used to call a router. Everything else can be provisioned.
Providers within the network-as-a-service market will be able to take on all -- or almost all -- of the network element of IT. We still recommend organizations retain institutional expertise, in the form of either internal or external consultants. It's possible multiple network-as-a-service providers may be required, particularly in international settings and where reliability is essential.
In the years to come, network as a service may evolve into something more intriguing. As 5G gains traction, mobile users may, in effect, be bringing in their own networks. The 5G features may reduce or eliminate much of the requirement for local networking. To that end, network as a service, in some form, is likely to become an element of most enterprises within just a few years.