3 Ways Containers and Microservices are Changing Network Visibility
Jun 25, 2021
So much of good IT management is about looking into the future. End users’ primary concern is right now—is the application they need for an urgent task available and performing well? Executives often focus on look-back data, such as variances against this years approved budget and trends across previous quarters. But ITs view needs to be on what’s ahead. What will your business need as it grows? How are technology advances changing the way your organization gets things done?
The use of containers and microservices is one area that seems to be on the horizon for many companies. However, the reality is that their use is growing, and many enterprises are farther down this path than they think.
In a 2021 survey, 65% of respondents said they use Kubernetes in production. This is a jump from 2020’s 59% response rate.
Whether you’re still dabbling with containerization or microservices have been part of your infrastructure for a while, your environment is already headed in a new direction. An expanding reliance on microservices and containers will create gaps in your network visibility that could sneak up and bite you, and the time to get out in front of these vulnerabilities is right now.
1. Fragmentation of Services
Containers’ effects on the distribution of services represents fragmentation—of ownership at the enterprise level over specific workloads, of operationalization of those workloads and of the way they’re broken into smaller chunks of infrastructure. It’s analogous in many ways to what SD-WAN is doing to network workloads, distributing a network presence and controlling it remotely via software to provide performance, scalability and security at a wide scale. The container is the software equivalent of that.
But while the advantages of containers and microservices are important in today’s highly dynamic landscape, they introduce some key challenges from a visibility standpoint. Where IT previously focused on knowing whether an application was up or down on a single server, the introduction of containers changes the picture. Now that same application could run across many containers, regions and hosted zones, through different clusters across the world. Distributing your workloads yields big benefits—services are closer to the end user, better performance, plus increased reliability and scalability through clusters that can quickly spin up or down. But you also need to be ready to tackle the problems that inherently come along with a new networking strategy that’s far more complex than the traditional hosting model.
2. Tooling & Infrastructure
You need to match that emerging complexity with corresponding maturity in your toolset to handle the new way of networking. It’s no longer enough to say that an end user connects to a router or MPLS that then connects to big web servers. That’s a very simple flow where problems can often be fixed by simply throwing more resources at them.
Microservices and containers introduce more complexity. The way the network routes across that path is different. You’ll be working with multiple network paths to multiple clouds, and multiple application sources in multiple containers. The challenge of maintaining visibility becomes more difficult, both in terms of what needs to be captured but also in the scale of what must be watched.
3. Support & Processes
There are two steps businesses should take to ensure they’re prepared for more vigorous visibility capabilities to support their microservices footprint.
1: Take a step back and understand that humans can’t manage everything that happens in a containerized environment. IT teams have already adopted tools to help them better manage application performance, such as log and server monitoring that provide awareness into how an application is performing.
2: It’s critical that you adopt a discipline—the processes as well as the tools—that allows you to maintain visibility at scale. It’s not possible to write a one-off script. You need automated deployment and management, visibility and monitoring, and all within a single pane of glass to truly understand how the network is performing.
Walking the full path from cloud monitoring to application monitoring to network monitoring to end users is complex and difficult. It requires targeted tooling and the ability to collect data, to correlate and enrich it, so you can turn it into something that’s actionable. An AIOps layer will enable you to ingest and process the data so it can then be delivered to the well-oiled machine that is your network operations center, or to your support team that has developed good workflow automation and visibility to remediate issues quickly.
Together, these steps will allow you to maintain short MTTR, to avoid long outages and to more quickly track down outages, even as your network complexity grows.