For years, enterprises have budgeted network management at a percentage of their total network spending, and for most of the last 15 years they've been squeezing management spending down to a smaller percentage every year. But starting about two years ago, this trend reversed, and in 2013 businesses at all levels budgeted more for network management tools than they had the prior year.
Could it be that (gasp!) network management is getting cool? According to enterprises, there are three factors driving a major change in their management vision, and driving more spending to boot.
The first factor is increased dependence on networking. For decades, information technology has been inveigling its way into the heart of business operations. In the old days, having a worker disconnected from the network might force rescheduling something, but it probably wouldn't have had an overall productivity impact unless the outage lasted for days. Today many workers are totally dependent on network-delivered information, to the point that their productivity falls instantly to zero if the network breaks.
We have a modern example of why this happens. Mobility empowers workers by letting them take their IT resources with them literally in the palm of their hand. Inevitably, that means that if you take away the connection to that worker because of a network failure, productivity will fall at a minimum to its earlier (connected but pre-mobility) levels. If worker practices were tuned to be dependent on the delivery of information to them via a mobile device, then their productivity might fall to zero. In an effort to gain a benefit, we're building a dependency, and that's been true all along with network-based information delivery.
The second factor is the increased complexity of network infrastructure. In the good old days, we had modems on client devices and a modem bank at the server, and people called in. Today we have routers, firewalls, NATs, DNS, DHCP, VPNs, application delivery controllers or load balancers...you get the picture. Not only that; every year, vendors try to improve their competitive differentiation and add to their profits by enabling new features. All this stuff has to be managed.
In my Fall user survey, a quarter of users said that managing their security products alone required as much effort as was required to manage their entire network just a decade ago. And who believes that networks of the future will be getting simpler? I asked users that question, and it probably won't surprise you to hear that nobody believed that would be the case, and, in fact, almost all users thought they'd be "significantly more complex" to manage five years from now.
Point three, our final point, is that we're growing network dependency faster (much faster) than we're growing network skills. Today, three-quarters of all enterprises say they have a shortage of skilled network professionals, when only a decade ago less than half said that was the case. Today, three-quarters of SMBs say that they effectively have no access to skilled network professionals at all--either because they can't pay them enough or because they just aren't present in the local labor pool.
Training people isn't going to work, either. Half the people SMBs train as certified professionals leave the company within 18 months for a higher-paying job, according to the survey. Even without formal training, SMBs have a hard time keeping anyone in a network support role for even two years.
Some see this as a big opportunity for outsourcing, but virtually all SMBs and most enterprises say that network support outsourcing is too expensive for them. In the last five years, spending on outsourcing has gone up about a quarter, while spending on management tools has doubled. And remember, outsourcing of anything is dependent on economy of scale, and that means that even network management outsourcing or complete network outsourcing is going to demand more effective management tools to reduce the cost of support.
How does that happen? Don't expect the "network management" term to suddenly acquire cachet. Instead, what we're seeing is a change in focus. Traditionally, network management systems have supported the network operations center (NOC) and the staff there. Today the problem is that there isn't anyone to staff the NOC, so instead, companies want something to replace it. Network management a la OpenView is going to turn into what might whimsically be called "NoView". Let automated processes run everything; that's the message.
This transition will have far-reaching impacts, and perhaps the most significant will be that it forces us to think in terms of application-as-a-service, meaning a bundling of network and IT resources into a pool that supports worker experience. "Network management" first becomes "service management" and then becomes a set of automated processes to sustain the level of experience that was required to meet the business case.
We can see this starting even now. We hear about "orchestration" of resource behaviors, and about "DevOps" and even about "MANO" (Management and Orchestration, the Network Functions Virtualization name for service automation). As soon as we can start to think about enterprise applications as "services," we can apply general principles of service automation and lifecycle management to them, and these are the two areas where we've proved we can shift the focus of network management from the NOC to robotic processes.
With enterprise literacy on these topics exploding, you can bet that vendors will be spinning tales there soon, and that's good. Old management may never be cool, but its offspring has a real shot.
Follow Tom Nolle on Google+!
Tom Nolle on Google+
