Trends in Remote Computing
by Jaap van Duijvenbode on November 16, 2017
What happens when big data, mobile computing, and IoT converge? You get edge computing -- and special data consolidation and analysis strategies to accompany it.
What Is Edge Computing?
Edge computing refers to the practice of performing data analytics and processing on or close to the remote devices -- like smartphones and IoT sensors -- that collect or generate data. In other words, data processing occurs on the edge of your infrastructure because it takes place on or near the peripheral devices within your network.
Edge computing is the opposite of uploading data to central servers and processing it there, which was the traditional practice.
The advantage of edge computing is that it avoids the potential network bottlenecks that can occur if you transfer large amounts of data between peripheral devices and central storage hubs. Edge computing can also help to reduce the overall size and, by extension, the cost of your infrastructure because it offloads work from servers to peripheral devices.
As remote computing grows ever more common as a result of the expansion of the IoT and greater use of mobile devices, edge computing is an essential architectural strategy for ensuring that data can be processed efficiently and quickly.
How Edge Computing Changes Storage Strategies
Making the most of edge computing requires more than simply moving workloads from central servers to the edge of your infrastructure. In order for the edge to be able to handle the workloads efficiently, you need to develop the right data management strategy to support edge computing.
Such a data management strategy should include several components:
- Data consolidation. In many cases, data processing within an edge computing architecture does not happen directly on peripheral devices. Your smartphone or IoT sensor may not have the resources needed to perform data analytics effectively. Instead, data is consolidated from the devices onto processing nodes that are in close proximity (from the perspective of the network -- meaning they are within a few hops) to the devices. Integrating data from diverse devices and transferring it to edge nodes for processing is, therefore, a crucial part of an effective edge computing strategy.
- Data selection. Some data (such as information about a failure that can be used for troubleshooting) needs to be transferred from the network edge to central servers. Other data can remain on the edge and be processed there. Selecting which data to send to which part of the network is essential for managing data effectively.
- Data mining. Making the most of data in any environment requires mining data for important insights. In edge computing, as in any type of computing architecture, leave no data source unexamined.
- Data security. In some respects, edge computing makes data security easier to manage because less data is transferred over the network. But that doesn't mean you can forget about data security. Securing data at rest and in motion, as well as ensuring the security of the devices that host and process your data, is just as important at the edge of your network as it is when dealing with central data storage servers.
In short, although edge computing leverages an unconventional type of architecture to make computing more efficient, edge computing does not mean compromising the thoroughness of data management.
With edge computing, you still need to build data lakes, consolidate data, mine data and secure data. The ways in which you work with data on the edge is different in some respects from data management within traditional computing environments, but the same core data principles and strategies apply.
TalonFAST provides the data management features you need to store, mine and secure data effectively in an edge computing architecture. Click here to learn more about the TalonFAST solution for data management.