Endpoint and Host security techniques have diverged. There used to be considerable similarity between the techniques and tools used to secure desktops, servers, and even networks. Desktops evolved to become Endpoints, as mobile devices proliferated and they were assembled into a collective of being in the category of not-a-server.
But as practitioners know, it isn’t all about the tech. Organizations changed, too: operations groups diverged into distinct endpoint ops and data center or server ops teams. At the same time security shifted to more often becoming about monitoring and ensuring security than by implementing it and operating it.
BYOD and mobile workers pushed Endpoint Protection Platforms (EPP) into new tasks, like encryption, application control and DLP. Server communication and composition changed, and servers became increasingly virtual and are highly replaceable through orchestration. Server security became very different and subdivided into unique types reflecting the different exposures between web servers, data servers, and others.
The future sees this trend continuing. In the near term, containers mean that intra-server security becomes more complex and a bigger task. Containers shouldn’t normally house data, but they handle it and therefor become a target, especially if the application security regime doesn’t include container security. The communication between containers and between containers and apps becomes a key point to embed security. But likely not the only place, as the algorithms inside apps and containers become a future target. In the mid term, serverless becomes the new milestone in a changing data center as to how servers evolve and present new security challenges. “Is the server secured?” becomes a degree more abstract and moves towards “Are the server(s) secured?”
Network security is under organizational pressure as network ops gets sometimes forgotten in the move to hybrid and cloud. Contrary to the belief that network security goes away, it becomes more critical as your data moves to new and odd places. Endpoints and servers still need securing, and so too does the connection to them. There isn’t enough space in this blog entry to cover all the security dynamics of cloud security.
Clearly, all this disruption and specialization has created very narrow silos. Meanwhile, attacks are going low and slow. Within these silos the challenge today is in spotting meaningful attacks but how fast the label of meaningful is assigned and then the time to resolution of an alert. And that is within a silo. As the whitespace between the security silos of endpoint, server, network and cloud expand the opportunity to pull together becomes more important. SIEM is a critical tool, but the pre-SIEM and post-SIEM phases in reducing alert resolution effort and time there is wealth of security-relevant information lying between these silos. Better glue between these silos can mean better pre-SIEM secfurity operations creating fewer alerts to resolve, what comes out of the SIEM is more relevant, and these resulting alerts can have a faster time to resolution. Too often we rely on the least-scalable resource to be this glue between the silos, our “meat computers” – our people. There are other security and tech silos as well I haven’t mentioned such as data, personnel, and data center.
The greatest challenges and opportunities in security present themselves when the organization and technology go orthogonal. As our technology, security, and organizations gets more specialized and more silo’d, putting more non-human security glue between these silos is a big opportunity. Specifically, recognizing that security happens across silos, even when security itself isn’t structured that way.