Enterprises today are increasingly turning to software-defined networking (SDN). The versatility of multi-cloud, paired with the freedoms granted by unshackling from a single infrastructure-as-a-service (IaaS) offering, has acted as enough of an incentive for enterprises to make the switch to centralised and more dynamic SDN environments.
Adoption shows no sign of slowing—according to IDC, the SD-WAN Infrastructure Market is poised to reach $4.5 billion in 2022—that’s a 40.4 percent compound annual growth rate within five years.
Despite its vast potential for the industry, implementation poses new problems. SDN environments require specific approaches and tools when it comes to monitoring. It’s in the best interest of all enterprises to establish these systems and processes from the outset. Doing so will help massively down the road, when SDN stops becoming the next big thing—and becomes the accepted status quo.
It’s easy for IT pros to become complacent in the face of the potential benefits of SDN. All pros should be conscious, however, that these networks are incredibly complex and present new and unique challenges that will need to be managed. Understanding how to deal with these challenges will be vital for IT pros looking to sufficiently monitor the networks of tomorrow.
No change, no gain
With adoption of SDN so high, it’s vital for enterprises to start weighing up how such big infrastructure changes are going to affect the day-to-day monitoring of network performance and security. These enterprises would be wrong to think that these changes can be bolted on without warranting any change. Maintaining existing monitoring systems and practices without audit will leave IT pros blind to the fatal holes their new SDN environments may have opened up in the network.
SDN is more than a simple rejig of the traditional IT infrastructure model—it changes the entire paradigm. SDN environments are intelligent and flexible, meaning services and devices can enter networks almost instantly, and, if the right tools and processes aren’t in place, often without IT knowing. This renders traditional daily security checks obsolete. SDN environments are constantly changing, so effective visibility will only be achieved through dynamic, real-time monitoring.
Multiple tunnels and layers also make the performance of SDN environments difficult to monitor. Basic bandwidth reporting is less effective than in legacy networks. In fact, focusing solely on measuring bandwidth can hide instances where an individual vendor’s cloud service is failing and causing performance issues in the entire network. A strong bandwidth reading on the control panel could still equate to connection issues for a frustrated end user.
More vendors, more problems
It’s clear then, traditional monitoring practices don’t cut the mustard in SDN environments.
Enterprises often initially stick with a single SD-WAN provider, but along the journey of building out their networks—different cloud providers are onboarded, and various parts of the network are automated using different vendors, and suddenly organisations are required to study three or more different APIs. It won’t be uncommon to have a different provider for on-premises SDN, SD-WAN, and the cloud.
Enterprises can’t be blamed for embracing these multi-vendor environments. Wholeheartedly committing to one vendor, obviously, causes vendor lock-in and means enterprises will be unable to bring in more suited or better priced vendors along their network building journey. This kind of approach also means enterprises can access best-of-breed services from each vendor, ensuring their networks are supported in the most efficient ways.
However, from a monitoring point of view, this can complicate things. Every vendor brings their own individual dashboard that provides visibility into only a very specific fraction of the network. This shatters the “single pane of glass” standard needed for optimal monitoring, leaving a disjointed assortment of dashboard shards in its wake.
Wading in fractured landscapes
This is why it’s so important that, from the outset, IT pros are looking to monitor SDN environments with vendor-neutral performance management and development platforms. A key part of this is ensuring performance monitoring contextualises the entire network, and the functionality of the vendors and cloud providers within it. Metrics such as log and flow analysis and deep packet inspection are much better suited to measure the performance of multi-vendor networks than basic WAN indicators.
The industry itself is also working to solve the challenges posed by multifaceted environments. Just this month, Oracle and Microsoft announced a new partnership that will see Oracle Cloud and Azure directly connect—enabling users to move workloads and data seamlessly between the two. Close collaboration between cloud service providers is becoming increasingly common—this can only improve the situation for IT pros, tasked with monitoring an increasingly complex web of hybrid and multi-cloud environments.
The interoperable dream
The potential hurdles in the monitoring of SDN environments all hinge on the fact that modern networks are disjointed and complicated.
The good news is that vendors are increasingly creating services in a standardised way, using consistent APIs based on open standards. This grants a new golden opportunity for developers to consume all these APIs to create single, unified monitoring layers for networks.
Standardisation is the ideal antidote to the difficulties of monitoring today’s complex SDN environments. With the sharp rise in popularity of APIs and open source amongst vendors of all sizes, this standardisation is increasingly becoming a reality.