Data gravity, the principle that as data gains volume, it also gains “mass” and draws elements of the overall software architecture closer to the site where data is generated or processed, has continued to drive modern architecture. This trend has forced applications to become more modular, enabling components of applications requiring low-latency or high-bandwidth access to data to be pushed to the network “edge.” In 2023, the various trends mentioned above played into this bigger trend, requiring hybrid cloud architectures to respond to the data needs of applications, like AI needing direct access to data at the edge, which often means on the manufacturing or process floor. As an example, the U.S. Army has been doing “edge computing” for years, calling it the “Tactical Network,” which allows for the rapid and reliable exchange of data at the point of need in contested environments.
In edge computing, an increase in data and processing outside the traditional data center creates potential risk to organizations, as reduced physical security, a limited compute footprint, lower cost expectations, and remote management are often compounded by a lack of IT personnel. Security concerns are on the rise in this environment, as centralized control and distributed execution is the name of the game. Modern advances in theory around Zero Trust Networking are making it harder to understand the security impact, but standards such as FIDO Device Onboard (FDO) are helping to ensure a secure future for edge computing. Implementations have matured a lot in 2023 and we expect the market to mature much more in 2024.
Cloud computing has two particularly valuable use cases: organizations wanting to deliver a product fast without building up or hiring infrastructure expertise, and organizations that want to elastically scale their capacity to deliver to end users and customers. Cloud was promised to radically change the way organizations run their infrastructure, and some bought into the promise of infinite connectivity, infinite compute, and infinite storage. But there is a cost for such things, and cloud vendor bills for storing the massive amounts of data being fed into AI models have elicited hints of buyer's remorse. Some organizations are starting to explore the possibility of repatriating that data, bringing it back into their own data centers and using on-premise systems and expertise to manage it. There is a real cost, however, to moving large amounts of data from the cloud, so the best answer is one that allows letting data remain wherever it makes the most sense and without restricting the important work you need to do with it and making the problem worse. In 2024, more organizations will hone their “return from the cloud” skills in a way that drives new innovations while optimizing the costs of working with data, transferred and not.
Interestingly, all these trends share a common thread: They are all about data gravity and the methods by which we move data or the applications that need it around the network. Technology systems in place today can barely keep up with the burgeoning storage and use of data worldwide, and we are continuously finding new ways to process, move, and store information. Sparks of intense interest, whether individual or commercial, will bring new ideas to bear on the problem sets we are encountering and new software-defined systems that will meaningfully shape businesses, the infrastructure we run on, and the networks that tie all of those systems together. 2024 is definitely going to be an exciting year for innovation!