In recent weeks, our blog has been focused on various critical components of today's IT service management suites. We have touched on IT Asset Management, Event Management, Job Scheduling and Automated Workflows and Orchestration.
Let’s now take a glimpse of where things are headed in Network Management, and the issues that continue to bedevil networks managers worldwide. It may seem like stating the obvious but if you ask any veteran IT ops professional, all of the tools and services that we have been exploring require the solid foundation of a stable network infrastructure.
The Challenges of Managing Complexity
Network managers suffer from the same issues that their peers in other areas of IT encounter when delivering services to end users:
- Too much data to analyze and correlate
- Multiple, disparate management and monitoring tools
- Mobile users with BYO smart devices accessing services not under IT’s control
- Lack of an overall, end-to-end view of the client experience
Even the most basic functionality – such as connecting to an enterprise network! – comprises a series of complex transactions, each of which can impact the experience. Connecting to Wi-Fi, authenticating to the network, obtaining an IP address, resolving a URL, getting routed to the appropriate domain – each of the things that happen under the covers that we take for granted are not trivial from a technology perspective. Often the only clue that the support team has is an angry message: "I can't get in."
The Wish List
Today's infrastructure – network, apps, premise, cloud – generates enormous quantities of data, logging and otherwise. Nowhere is this truer than in the network itself. And one thing that everyone agrees on is that your network folks can’t sit around staring at data 24 hours a day.
To manage this muddled matrix, network managers have been forced to use a mixture of discrete vendor and homegrown tools, each of which provides some sort of view into specific parts of the network. Is it network hardware? Is it an issue with one of the dozens of protocols handling the various transactions? How many users are downstream?
The tools a network engineer might use to troubleshoot WAN or LAN issues are much different that the tools the applications or systems admin might use – and none of them tell the entire story. Different monitoring tools show the blind man different parts of the elephant.
Just plowing through the logs from a single server to figure out why one user can't accomplish what needs to get done can require a vast amount of time. Many of us have experienced this firsthand.
What network managers are really looking for is a heterogeneous platform that can be used to provide specific insights across the wired and wireless network, the wide area network and application services. They need tools that automate the learning process of how the myriad user devices actually behave with every part of the network. The need a single source of truth that examines and quantifies every aspect of the client experience.
Light at the End of the Tunnel
A few approaches are emerging that look promising. As with many areas of technology, these tools leverage recent advances in machine learning and big data analytics and marry them with cloud computing.
Network traffic – both wired and wireless – can now be delivered to a server that crunches the data, looking for problematic trends and patterns impacting user connectivity. These user performance management tools typically measure and analyze every client network transaction to learn how the network, services and applications are behaving with all client devices, providing a holistic view that exposes issues.
If there’s a Wi-Fi coverage problem in a specific location, a DNS connectivity issue for a given group of clients or an application response time problem, these solutions will identify it and flag it for remediation.
The other major shift underway is where the machine learning tool looks for patterns impacting user connectivity. Network management tools are evolving from being fault based to performance based. Fault-based monitoring was crucial in the days before infrastructure was built to be as redundant as it is today. Most modern networks can survive a switch or Wi-Fi access point failure without severely impacting the overall user base.
Where Things Are Headed: User Performance
A bigger problem is managing user performance. Often users call in about a certain application not working well, but when the engineer looks at the dashboard, everything is green. Performance problems are much harder to diagnose and can kill employee productivity.
The problem with traditional network management is that by definition, it provides a bottom-up view of the world where user experience is inferred. User performance management is where things are headed. It’s more top down, so the underlying technology and dependencies are understood.
That means if any given network component gets overloaded, the network operations team can quickly understand which applications and users will be impacted.
Clearly, this type of functionality requires close integration with the rest of the ITSM suite, but the benefits are tremendous. Struggling with hard-to-diagnose network issues can cripple productivity company wide.