IT teams tasked with centralizing network performance management and security are challenged to codify incompatible systems, heterogenous infrastructure elements and disparate vendor tools designed to provide only a partial view into a given part of the network.
Every piece of the IT organization has its own little area to deal with. The wireless team manages access points. The network team manages switches, routers and network services. The server team manages servers. The security team manages firewalls, network access controls and policies. The WAN team looks after the utilization of broadband connections. You don’t typically play outside your area and work with the other teams unless you have to. If problems are traced out of your area, then it’s inevitably their fault.
By eliminating IT silos, organizations can then enable new intelligence, more effective decision-making, and informed automation to optimize network performance management (NPM).
That is to say, if siloed walls are broken down and the synergies between tasks and technologies can be better coordinated everyone benefits. The question is how to get there without going broke or crazy.
According to a new research report from Enterprise Management Associates, IT operations teams are struggling to assemble an effective NPM strategy upon which everyone agrees. The report asked hundreds of business how they are looking to address gaining a tighter grip on managing the performance of their networks across different IT groups.
Topping the list of technical initiatives that influencing their IT NPM decisions were IoT and IT analytics (see chart below).
Technology initiatives most responsible for driving NPM strategies
Context is everything
Nearly all Infrastructure vendors provide some sort of analytics tool to manage and monitor what they sell – Wi-Fi troubleshooting tools, mobile device management (MDM) systems, WAN utilization analytics, application performance monitoring (APM). There’s something for every IT faction. Any technology or tool that cuts across the network to deliver actionable analytics for different IT constituents is a big win for the business. The problem is really about context.
For instance, users often blame Wi-Fi when they can’t connect, or network performance is poor. Successful client connections not only rely on associating with WI-FI access points but also authenticating users, obtaining IP addresses, resolving domain requests, receiving timely application responses and often traversing WAN links to reach cloud-based applications. So where is best place to start looking to find potential performance or security problems?
Such situations naturally produce isolated IT silos and finger pointing when things go wrong. Many of these silos are tied to business processes sprinkled around the company, so transitioning to a centralized IT model can not only be time-consuming and frustrating, but also very expensive.
Enterprise IoT only makes things worse
Exacerbating the silo effect is a flood of new Wi-Fi-connected IoT devices, critical to line of business objectives, now littering enterprise networks.
This year, IoT devices will outnumber the world’s human population for the first time. By 2020 it’s widely expected that there will be nearly three-times more connected devices (20.4B) than people on the planet.
Who is responsible for IoT device management, performance and security? Is the line of business who purchased the system? Is it the Wi-Fi engineers? The network team? The security staff? The answer remains unclear from organization to organization and an ever- increasing point of contention.
With more pressure, IT, security and networking staff must find a way to codify infrastructure management and security operations. Winning organizations and teams will break down these silos. But how?
AI and Analytics: A perfect marriage?
The answer is in the marriage of analytics and Ai to provide so-called operational assurance. The combination of these technologies now gives service desk, security, Wi-Fi, network, WAN and application teams across the enterprise access to a single source of IT truth bolstered by quantifiable data analysis.
New infrastructure management solutions, dubbed AIOps edge platforms, are embracing the analysis of infrastructure, device and data analysis across the entire network stack from a strictly vendor-agnostic perspective.
Instead of relying on discrete vendor tools, manual inspection and human interpretation, these platforms automate critical IT processes by ingesting large volumes of different tata types, constantly determining the normal behavior of virtually every dimension of the network (see figure below).
How it all works
These new systems, deployed out of band, use big data analytics and cloud computing to baseline application, Wi-Fi, network service, device and WAN performance. All network packets are spanned to a collector that also talks to other network elements such as WLAN controllers, AAA systems and routers using standard protocols and APIs. This information is used to glean and correlate details about how any client device is functioning with other services, devices and applications across the entire network.
When a device or device group deviates from “normal” desired behavior that would negatively impact the user experience, IT staff are automatically alerted to the logical location of the incident across the stack, client details, the blast radius of the problem (i.e. was it a single client issue or a network-wide issue) along with recommendations on best to remediate it.
For example, when a new IoT device is attached to the network, these systems can quickly identify and classify the device – then begin watching how it is behaving with other parts of the network. If the device is trying to access a suspect host destination that other devices like it aren’t or is showing up on a VLAN or SSID it shouldn’t be, IT staff is immediately notified and can then take immediate action to segment, micro-segment or quarantine the culprit clients.
The beauty of these AI-based platforms is that they aren’t tied to any one vendor and are using real network data and standardized protocols such as SNMP, NetFlow, syslog, and APIs to do their job. As a result, these systems become immensely useful to the entire IT organization by providing detailed insights from infrastructure data to different groups who want to use it for their own purposes.
Now IT teams can better collaborate, remediate issues and get to answers to complex questions without having to wade through volumes of log, client and network data. And because each IT discipline is looking at the same data, just from a different perspective, traditional IT silos naturally begin to diminish.
Ultimately, IT organizations need to start breaking down traditional silos by implementing these new technologies and AI-based approaches that leverage the one thing that (almost) everyone agrees doesn’t lie: network data.
By democratizing the interpretation of network data across the IT organizations using a new era of AI and analytics, traditional IT silos begin to blur, and cross-function collaboration becomes something IT professionals can embrace.
