Mark likes to joke with me about network management – based on my past history (and what he believes is my love affair with it).  Over the years, he has convinced me that it is appropriate to refer to network management tools as “toys in the kids closet” after a major holiday– lots of once-used, never-opened, tried-and-set-aside, toys strewn about and discarded.  They all seemed like a good idea at the point of sale but never worked as advertised and therefore were thrown back into the messy closet.  Over the past 10 years or so, I have often been reminded that enterprise IT managers will get around to network management once they are done with the important things in their network. 

But . . . things may be changing.

While working with a couple of network and service management startup companies recently, I had the chance to speak to dozens of enterprise IT managers and I learned some interesting things about the state of network management products in the enterprise.  Specifically, IT managers are finally ready to spring clean that closet that they always said they would.  These managers are driven by the increased focus on costs and proven ROI to go back and determine just how (and if) they can make the tools that they already have work for them.

So, while this closet-cleaning is well underway, let’s take a look at what is getting a great deal of their attention. 

What’s hot?:

The issues that are hot in network and systems management have a lot to do with what’s already been done and is working.  Element managers and performance management products have been working very well for many years in the corporate enterprise and they deserve a lot of credit for doing what they promise.  Other areas of increased interest are…

Security – there is continued interest in security both as a result of 9/11 and the ongoing effort to externalize enterprise applications to business partners and customers.  But what’s changed is the realization that firewalls aren’t enough.  While the jury is still out in terms of effectiveness, Intrusion Detection systems are being investigated as well as tools that maintain data trails of break-ins for the purpose of satisfying a security audit.  What’s still missing?  Security products that are application and user-aware and can distinguish normal behavior from egregious behavior at the application level, and some form of system to bring all the security information together into one composite view of the security “health” of the corporate system (applications, data, and network).

Fault correlation – for a number of years, frameworks and other tools have been functioning as trap receivers and alarm monitors – essentially showing what is broken within the network components.  As SNMP traps have become pervasive, IT managers have come to realize that they need to find a way to wade through the literally thousands of traps that they receive – many of them useless – to figure out what is really wrong.  I think most of these managers have come to realize that a system 100% effective in determining the root cause is a bit like searching for the “holy grail.”  But instead of giving up, they are placing a renewed emphasis on fault correlators as a way to get to the root cause for about 80% -90% of the faults – a more attainable goal.  This has a huge benefit for their overburdened staff – especially the gurus - and allows them to create an “infrastructure cockpit” where they can see the overall health of the enterprise.

Storage and Data Management – the past couple of years has witnessed the shift from legacy storage to networked storage.  This has led to a significant rise in importance of storage and data management.  Everything from SAN infrastructure, to logical partitioning of data, to data protection and to simple NAS storage, all have to be managed – and often by a group of people who have historically managed networks.  A real potential for culture clash here – but fortunately, there are a number of existing and emerging tools that make it possible to realize the goal of storage as a utility to the applications and systems in the enterprise.  This is a high-priority item for most enterprises.

Controlling costs – this has never been more in focus.  IT budgets are increasingly tightened and measured against the revenue of the company.  It’s the way to understand the budget expenditures as compared to other companies.  This has led to awareness that IT managers don’t always know the service level impact of a budget cut.  To some extent this has led to increased emphasis on planning tools and fault management systems and pseudo asset managers to see what is least used or what can be done without.  But the reality of this is there is an opportunity for a high-level service level management product that ties in faults, performance, and criticality to the business processes.  This doesn’t exist yet – but IT managers are starting to look for it. 

What’s not?:

It’s funny how some things that seemed so inevitable only a couple of years ago could be so “old-hat” today.  It’s not because they were bad ideas then, some of the fundamental assumptions changed.  Here are a couple of obvious ones that I didn’t hear anything about in my recent conversations.

QoS – Quality of Service, one of the more common buzzwords has been put on the back burner.  Not because it is unimportant but more because the bandwidth glut in the enterprise and service provider networks has made traffic prioritization less important that it used to be.  There are still important things to be done –like allocation of bandwidth to critical, response time sensitive applications.  But the overall issue of QoS has been largely relegated to the network access where bandwidth is not as plentiful – if the subject is being discussed at all.

Policy Management – about three years ago it was hard to have a management conversation without discussing your position on Policy Management or Directory Enabled Networks or some such derivation.  Now, I don’t mean to belittle the effort, but not everything fits neatly into policies – at least practically.  Policy management has been (rightfully) broken down into more digestible pieces.  Security policy? Yes.  Application usage policy?  No.  User Access Policy?  Yes.  Traffic Prioritization Policy?  No.  I guess the over-riding issue of policy management will have to wait a few more years until the IT Managers get their closets cleaned.  Maybe then they’ll have the fundamental underpinnings (and lack of clutter) to get a meaningful policy management system thought through and in place.

As IT Managers search through their closets for tools to help them squeeze more out of their networks and applications, they will find they already have the right tools in many cases, but in other cases they will find the toys in the closet to be broken, outdated, or too juvenile.  This opens up the opportunity for new management products to fill the needs of today –sort of like the Thrivers and Survivors Mark talked about in last month’s musing.

Next month, I’ll take some time to discuss some of the more interesting ideas that I have seen coming as a result of the unsatisfied needs of customers.

So, Mark, my long-standing prediction about the importance of network management may finally be realized – sort of.  Spring cleaning is in – and even though it might take the majority of 2002 to get done (I guess the closet did get a little messy over the years, didn’t it?) we’ll come out of it better off in the enterprise and with some clearer opportunities for the startups.