Blog: Do You Have a Strategy for Storing Data from Your Network’s Edge?

As the IT environment evolves, meeting storage needs demands new strategies. In the old days, most enterprise data was generated on-site and centrally stored in the data center. The edge of the network was defined by firewalls and was limited to the physical confines of offices. The network edge included some branch offices and a handful of enterprise-owned mobile devices. Storage at these sites was handled by SAN, NAS, and DAS boxes.

How Storage Has Changed With IoT?

This storage picture has changed rapidly and dramatically.  The proliferation of laptops, mobile phones, and other devices, often owned by employees, forced a rethink about where and what the network edge is. This was accelerated by the Covid pandemic, which kept employees out of offices and in their homes.  And, of course, there’s the tsunami wave of IoT devices, which is not just impacting many industries but also redefining them.

Gartner anticipates that 75% of enterprise-generated data will be generated outside data centers or the cloud by 2025. And that’s also where they will be processed.

Data created at remote offices is generally stored locally and backed up over the WAN to the data center or cloud. Much of these data is unstructured and not time-sensitive. No one will care or even notice if a PowerPoint presentation incurs some latency when being uploaded.

IoT Data Storage Management

IoT devices, however, create increasingly vast amounts of data, much of which is time-sensitive. Self-driving automobiles and trucks, for example, must know instantly when to brake for red lights. Moreover, not all IoT-generated data are essential. Generally, a heart monitor needs to issue alerts only when a patient’s thresholds are exceeded. Video surveillance footage needs to be saved for specified periods of time, but every frame doesn’t have to be preserved.

The bottom line is saving all this data centrally in the data center or cloud will put far too much pressure on wide-area links and result in latency and reliability issues. Additionally, uploading all raw data would be very costly.

You must devise strategies to store edge data locally where they’re generated, and forward only the data you need to the data center. Your solutions will be determined by your applications and how much bandwidth each use case requires. You’ll need healthy throughput to collect today’s high-definition surveillance video, but not so much for many manufacturing or medical devices.

 store edge data locally

How you store IoT data will differ from how you store data created by employees working from home. Additionally, you must consider security and compliance with remote storage. Remote storage will be difficult to scale so you need to anticipate your needs going forward. Also, your remote storage solutions must be robust and many will need to withstand hostile environmental conditions like temperature variations, dust, and vibrations. Finally, remote storage is more difficult to maintain and manage than data stored at the data center.

Role of SSDs in Data Storage Management

Solid-state drives will be appropriate for many use cases because of their durability and ease of maintenance, but low-cost spinning disks can still play a role. Large data streams like surveillance video can leverage local SAN, NAS servers that provide economical storage as well as applications that trim and back up the data stream. Consider IoT gateways, small servers in proximity to your data sources in the field that quickly process and analyze data streams and can forward just the results to the data center. They can reduce data collection expenses.

The need for storage at the edge is many, as are available solutions. Enterprises that cost-effectively collect, process, and analyze their remotely produced data will extract full value from them and gain competitive business advantages.