For federal agencies today, processing the massive amounts of data collected on a daily basis provides an unprecedented opportunity. The insights gained from analyzing this data can change the way we deliver citizen services on every front. But, there is a catch—as the amount of data collected skyrockets, quickly and effectively processing that data while keeping costs down becomes more resource-intensive and daunting.
As agencies continue to refine their approach to data collection and storage, they are finding that their strategies cannot rely on the public cloud alone. In order to scale computational power to match the potential of ever-expanding collection capabilities, agencies need to consider a comprehensive approach to data storage and analysis. Edge computing—a method of storing and processing data closer to where it’s created rather than deep in the cloud—is a key component to this approach. Here are some important considerations when deciding how to best use edge computing in conjunction with public cloud.
The Move to the Edge
There was a time when the industry believed it was moving towards an entirely cloud-based infrastructure. But as the transition got underway, it quickly became apparent that such a move would be so expensive and impractical as to make it ultimately impossible. While the price of storage has dropped, the cost of maintenance/analysis and the high latency of cloud-based computing has made many IT professionals turn towards the intelligent edge to help improve response time and lower overhead.
Edge computing aims to process data closer to the point of collection. The key benefit of analyzing data where it resides is reduced latency. It would be inefficient if an application being used by an agency in Washington D.C. had to call on data stored in Los Angeles every time it needed to perform an action, and impossible for applications operating with a spotty network or no connection at all. Also, developing analytical models for segmented data is much easier than parsing through massive data sets in a cloud.
But not all data can or should be dealt with on the edge. The need to avoid latency can be balanced with the benefits of the cloud to form an effective overall strategy with a few considerations.
Where to Begin
Most agencies make use of a wide variety of applications that are called on with different frequencies—some may be used by all employees every hour of the workday, while other, less critical, applications may only be touched once a week. Highly utilized applications are prime candidates for edge computing. The reduction in latency for something like an intake form, for example, will be highly impactful, especially on a large scale. Something less-frequently needed, however, like data collected for archival purposes or for monthly reporting, can often make use of the cloud, since latency doesn’t pose a fundamental risk to usability and performance.
In addition to utilization, agencies should consider the sensitivity of the information being processed. Sending highly confidential information across a network opens up vulnerabilities. While the cloud can be secure enough for many use-cases, keeping data processing for mission-critical information as close to the point of collection as possible will reduce risks.
Of course, when optimizing performance, agencies need to consider the preexisting state of applications as well as the current bandwidth of their networks. It is important to weigh the relative impact of incorporating edge processing for one application over another. Some applications are not built to have quick response times, and moving them from cloud to edge won’t provide enough benefits to make it worth the effort.
Many agencies have seen first-hand the benefits of moving to the cloud. Others may have been surprised by the costs associated with a consolidated cloud or found the potential risk and latency prohibitive. Either way, reorganizing how we approach data collection, processing, and storage is top of mind as agencies reexamine the role a cloud strategy will play in their IT infrastructures.
The intelligent edge can be an important part of that strategy, mitigating some of the associated risks, costs, and latency that would accompany a cloud-centric architecture. As technology develops in the direction of IoT, advances in machine learning have made the intelligent edge a rapidly developing frontier for both consumer and enterprise applications. Getting a handle on how to make the most of those advances will save agencies time and money and go a long way towards helping them accomplish their mission and increase the value of the resources they use every day.