Go to primary content Go to footer
Marines wading ashore

The Silver Lining to Military Cloud Computing

When exploring the military deployment of enterprise software distribution, cloud computing offers some key advantages.

While civilian and commercial computing operations are increasingly switching to cloud computing to manage software and provide data resilience, the case is different for government and military operators. Within a defence environment, cloud computing can come with a number of concerns from operators. These can include issues such as data sovereignty, cost, and vendor lock, when compared to standard on-premises setups.  

Armed Soldiers On Hill In Sunset

Furthermore, cloud computer platforms and data centres are not only a capability that is provided by commercial providers. Clouds can be created through data centres that are wholly owned and operated by a government or military, or ring-fenced to geographical locations to ensure data sovereignty.

Government requirements – both civil and military – for cloud computing come with some key similarities, ranging from data protection and security, through to the ability to deal with Big Data and leverage Artificial Intelligence (AI) technologies.

Despite these concerns, cloud computing offers distinct advantages for military and government operations. By leveraging secure and resilient cloud solutions, defence organisations can enhance efficiency, improve data accessibility, and strengthen cybersecurity. 

Here are four key benefits that make cloud adoption a strategic asset in the defence sector.

Rapid deployment and modularity

Hosting the core build of the software in a cloud environment allows for easier deployment and upgrading of capability for users within an enterprise environment. New versions can be more readily implemented when the software is based away from the user’s device, as well as the delivery of patches and hot-fixes, helping to improve functionality and security. This reduces the requirements to take a device-by-device approach to upgrades that can take time to implement, lead to uneven version distribution, as well as slower rolling out of add-on capabilities.

Cloud hosting, when utilising architectures such as Kubernetes and Docker, also make it simpler to deliver add-ons to users. As users may need access to additional features based on changing mission requirements, supporting access to add-ons can take advantage of pre-existing, developer-created linkages to deliver new functionality rapidly. Similarly, deprecating user access to add-ons as their mission necessities or licence requires is made easier through cloud management tools.

Scalability

As more data is being generated around the battlefield, from sensor systems to analytical reports, the need for it to be communicated to users at all levels is correspondingly increased. Cloud computing gives commanders the ability to receive this data and scale it according to their missionrequirements and information needs, with the fusing of data taking place in data centres that have a much higher computing power capacity. This scalability gives a tactical commander the ability to see much further than their immediate area of operations, and gain a greater comprehension of their part in the wider battlespace – thereby empowering their decision-making and appreciation of overarching strategic requirements.

Increased security and resilience

Cloud computing technology offers increased security and resilience for a disparate userbase. The dispersed nature of the technology helps to mitigate the risks around data loss, with an architecture that helps to combat cyberattacks, as well as controlling access to approved users, and more.

Housing large datasets, system processes, and core software in a cloud environment provides a greater degree of data resilience than processing data on a deployed server can give. Data warehouses in multiple locations means that if one centre is taken offline, then the other alternates can support continued processes as part of a PACE plan. Synchronising data across multiple data sites provides redundancy against physical and cyberattacks, while also facilitating scalable data delivery to users. 

The role of multiple sites with complex architectures also gives resilience from more routine cyberattack vectors that can be delivered against single-server systems. Attacks such as distributed denial of service (DDOS), and the maintenance of formalised intrusion detection and prevention systems (IDPS), help to provide more resilience for users. Formal certification requirements, such as ISO 27001, also help to continually improve the cybersecurity of a cloud computing environment for the benefit of all users.

Zero-trust architecture employed by cloud computing systems means that constant authentication and authorisation is conducted. This is required to maintain access to the requisite operational resources for functionality. This also means that cloud administrators have the ability to unenroll devices remotely to ensure operational security among users.

Offboard memory-intensive functions

Utilising cloud computing for military users means that more advanced technology can be employed by users closer to the tactical edge. As functions such as sensor fusion, geospatial processing, and AI become expected norms for end users, local devices may not be able to deliver. This can be due to device limitations such as memory capacity and trade-offs around ruggedisation, Graphics Processing Unit (GPU) and Central Processing Units (CPU) requirements, and power consumption.

multidomain-soldiers-and-helicopter

For mobile devices, such as tablets and mobile phones, the computing power requirements for deploying AI technology locally is burdensome. While commercial technology may allow for local AI operations, delivering this capability to a ruggedised military device may not be possible due to higher requirements around survivability, durability, and communications systems. Deploying a local AI model on a mobile military device could work from a memory perspective, but running a localised model can cause significant performance issues for a device in addition to requiring large amounts of processing speed and power. GPUs are preferred over CPUs for AI work as they allow for parallel processing, faster matrix computations, and energy efficiency. Integrating GPUs onto a mobile device is difficult due to the memory requirements, form factor limitations, and heat generation. 

In pushing the processing requirements for an AI model to the cloud, memory bandwidth on a local device is increased, response time for both general device functions and AI tasks are greatly increased, and specialised processing and memory hardware is offboarded from the local device.