Can Service Meshes Help Bring Legacy Government Applications into Zero Trust?

Zero trust remains a top priority for government agencies, although one that is difficult to achieve, because it requires a near total shift away from traditional perimeter security mindsets. By contrast, zero trust relies on strong identity management, constant user verification, least privilege networking and several other core components. Because of the level of change required, agencies have been slow to embrace zero trust, even though the Biden Administration’s Cybersecurity Executive Order issued in May 2021 requires it as a core component to cybersecurity.

The Office of Management and Budget has recently added more urgency to zero trust in government, issuing a memorandum to all federal civilian agencies in July that moves zero trust and IT modernization to the top of their priority list.

I recently moderated a panel discussion on zero trust with government officials from the GSA, CISA and DOD, alongside industry professionals, about the challenges of this new method of cybersecurity. When asked about the biggest of such challenges, legacy systems in government were one of the top concerns.

In government, legacy systems are not always comprised of aging mainframes from the 1980s or computers that still use floppy drives. There are many systems in government that perform their functions perfectly well with zero errors or disruption, but which might not be advanced enough to support modern applications, cloud deployments, complex networking or third-party programs. Those systems would be considered legacy, even though they may only be a few years old, and replacing them might cause downtime, disruptions or a long period of inefficiency if upgraded or replaced.

There simply has never been an impetus for replacing many of the legacy systems in government. Budgets are always tight, and if a system is doing its job, then replacing it does not make a lot of sense. However, many of those legacy systems were never designed to operate in a zero trust environment, which will increasingly become a requirement in the future.

I talked about this problem with some of the experts I met at different events, and a possible solution was suggested that I had never before considered. It may be possible to use a service mesh to connect legacy systems to a zero trust environment. This would enable legacy applications and systems to connect with a modern zero trust framework without having to replace the legacy system itself, and without having to add to its—likely limited—core capabilities.

What is a service mesh? 

A service mesh is most often employed in a containerized cloud environment to help manage how microservices interact. Most modern applications are built with services embedded inside them so that they know how to interact with others. However, in extremely busy environments, those hard-coded requests might get overloaded, which is why much of that traffic can be offloaded to a service mesh, which exists outside of the applications themselves. In a microservices environment, a service mesh allows developers to program core functionality into their applications while leaving the networking components to the external mesh. As Red Hat explains it, this lets developers change the networking components, program interactions and connectivity rules without touching the core microservices themselves, by simply updating the service mesh.

So, in general, a service mesh is a pre-configured application service layer that allows services to talk to each other, sharing data and consistency across an application lifecycle. You can almost think of it as the public transit network of a large city in that it connects all of the city’s amenities and services without directly affecting them. For example, if more people need to go downtown, they can be accommodated by adding more trains or buses, without having to do things like adding more parking spaces to the amenities themselves.

Each service mesh is comprised of two main components, the control plane and the data plane. The data plane attaches itself to applications and microservices using proxies, talking with them and issuing commands to and from the control plane, which handles all the actual networking and routing of those requests. Today, most service mesh deployments only work in containerized environments, and most of them tap into Kubernetes orchestration. The open source Istio service mesh is probably the most well-known example being used today, although there are others.

Tapping into a service mesh to help legacy applications

The idea which was presented to me by some of the zero trust experts is to use a service mesh to connect legacy equipment and applications in government to modern applications. This would allow agencies to continue to use their legacy applications like they always have. But the data layer of the service mesh would use a proxy to connect the legacy application to a more modern network. Essentially, it would add modern capabilities to a system that could not natively support them, without having to modify the application itself.

Once that was accomplished, the control layer of the service mesh could connect legacy platforms to any modern application, like an identity management tool being used as part of a zero trust deployment.

It does not seem like many companies are yet taking this approach. HashiCorp seems to be the leader in this field. They have a detailed video explaining how they are working on using service mesh technology to enable legacy gear to migrate into a zero trust environment. And Karsun Solutions is also working in this area with a pilot program that demonstrated that this new concept could actually work well, especially for agencies that have a lot of legacy gear, but which may not have a large enough budget to rip and replace it all as part of their move to zero trust.

Whether or not using a service mesh to connect legacy applications and gear in government with the modern systems that will provide zero trust is still an open question. But it’s an exciting area of study right now that may provide a better solution than trying to hunt down and replace every single legacy application in government. Even as an interim solution, it shows a lot of promise. 

John Breeden II is an award-winning journalist and reviewer with over 20 years of experience covering technology. He is the CEO of the Tech Writers Bureau, a group that creates technological thought leadership content for organizations of all sizes. Twitter: @LabGuys