Why the military needs a dynamic network infrastructure

Skirmishes are won and lost because of dynamic and accurate decision-making based on near real-time access to critical information. That information can come from a satellite, battlefield sensor or other edge device. Through artificial intelligence and machine learning, information becomes insights, providing troops with a strategic advantage over adversaries. However, warfighters will only gain a true advantage if the network infrastructure used to relay these insights is software defined, dynamic and secure.

A dynamic network can do multiple things well, resulting in a more informed and prepared military. It can facilitate real-time data processing at the edge while simultaneously supporting deeper analysis at the core, all while minimizing latency and bottlenecks. It can autonomously reduce the attack surface and minimize risk. It can also respond to limited or reduced communications environments and be ready to burst the data when connectivity is restored or re-established.

Let’s take a look at how a software-defined, dynamic network in theater can lead to faster and more secure communications and information gathering for the U.S. military.

What is a dynamic network?

To understand the capabilities and benefits of a dynamic network, it helps to define what it is. A dynamic network is software defined and has the flexibility to change and adjust based on different needs and circumstances. It can not only support cloud-scale operations on the back end, but it empowers edge and core compute capabilities to provide critical real-time intelligence. Along with API-driven software-defined networking, policies and orchestration enable dynamic networks to respond to real-time events and conditions as well as the changing needs of the applications and workloads they support. In the future, machine learning and AI will automate the policies and drive ever more efficient and timely dynamism in the networks and the services they provide.

For example, in a combat situation, warfighters may be receiving information from edge sensors telling them of enemy troops ahead. Sending that intelligence requires a significant amount of compressed data to be transmitted very quickly and securely. That requires a dynamic network capable of bringing the necessary compute and storage resources to the edge, facilitating an immediate response to a real-time event.

Another example is temporal data analysis at the edge. In this instance, critical data can be analyzed at each edge device. Using the power of these combined devices, warfighters can receive a more accurate picture of whatever is happening across their field of operations. Initial inferencing takes place at the edge, providing immediate feedback, while additional data is sent back to a core data center for cross-sensor analysis. A dynamic network delivers the inferencing models and compute resources virtually to different core or edge compute nodes to allow warfighters to respond in real time with the correct intelligence.

How does a dynamic network support data security and mitigate threats?

It is critical the military reduce any potential cybersecurity attack surfaces. As cyberattacks from enemy nation states grow in sophistication, the network must be nimble enough to detect and respond to anomalies before they become problems.

For example, sensors on and around the network may pick up anomalous signals and events that could be precursors of a cyberattack. Armed with this information, the network can spin up additional sensors to track these anomalies and, in tandem, throttle network services and introduce additional controls and gates as necessary. In short, with a dynamic network, the network configuration can change and respond to meet warfighters’ current state of threat or risk level.

Having a network that can automatically adjust services in response to attacks that are in process is also important. In a DDoS attack, for example, a network powered by machine learning can “learn” to automatically configure itself to move to a new IP address that is not being targeted or spin up new resources to handle increased loads. The network infrastructure itself can automatically change to successfully respond to the attack, all without the need for human intervention.

What’s next for dynamic networks and battlefield edge processing?

Field operations often take place in constrained environments -- remote locations that provide troops with very little network capacity. This can be challenging given the number of sensors and the volume of data collected in modern military operations. The network must be able to anticipate the capacity necessary for each of those sensors to perform effectively and minimize latency.

In this case, pre-provisioning for each sensor becomes incredibly important. AI requires constantly shifting infrastructure capabilities to meet demand for inferencing and processing as it occurs. One set of sensors may start tracking an object, then the tracking gets passed off to the next set of sensors as the object continues to move. The ability to anticipate which set of sensors will need what amount of capacity at any given time will prove to be a game changer that will lead to even faster and more accurate information gathering.

Military networks are not quite there yet, but they likely will be within the next few years. Dynamic networks will power this future transformation, just as they will power the kinetic battlefield of tomorrow.