The "Intelligent Edge" is an App Architecture Challenge
by Brad Johnson, on Oct 14, 2019 12:34:19 PM
Last week, Microsoft announced their Azure Data Box Edge computing appliance. While it may seem like a pure edge hardware play, CEO Satya Nadella positioned the product as part of Azure’s wider edge and cloud strategy, with a vision for edge and cloud applications working seamlessly together. “All this is being driven by this new tech paradigm that we describe as the intelligent cloud and the intelligent edge,” Nadella elaborated during his address at the Microsoft Government Leaders Summit in Washington, DC.
Nadella envisions future applications that will construct cohesive computational fabrics of all edge and cloud services. These edge/cloud applications will be further enabled by the use of AI innovations. Nadella elaborates that “of course when you have that rich computational fabric, one of the things that you can do is create this new asset, which is data and AI. There is not going to be a single application, a single experience that you are going to build, that is not going to be driven by AI, and that means you have to really have the ability to reason over large amounts of data to create that AI.”
It may seem like an intractable problem, consuming massive amounts of data to be able to train AI/ML. But the real challenge is how to build real-time applications that make use of these AI/ML tools to solve real-world problems. In order to confront this challenge, it’s necessary to shift from a big data architecture mindset to thinking about real-time distributed architectures, like the open source swimOS platform and our DataFabric enterprise suite.
Architecting Software for the Intelligent Edge
It’s certainly important to select the right hardware solutions for your edge computing use case, but it’s even more important to ensure you’ve selected the right software strategy and partners. In earlier posts, I’ve discussed at length the challenges of building IoT and edge computing applications using a centralized big data architecture. Edge computing requires making use of compute and storage resources in local environments, and it’s mission critical to create a cohesive mesh of edge and cloud resources.
There’s plenty of buzz around service meshes, data fabrics, and IoT platforms and it can be difficult to discern whether a specific solution is a good fit or an unnecessary layer of complexity without a real benefit. The challenge remains building software which can seamlessly navigate heterogeneous, dynamic hardware AND software environments. While market leaders like Microsoft and AWS provide solutions to streamline deployment and management of IoT and edge applications, few solutions address developers needs to building efficient, performant applications for edge environments.
How to Plan Your Intelligent Edge Architecture
The planning stages of your intelligent edge applications are the most critical for ensuring a successful project. By understanding potential challenges and pitfalls you may encounter, you can prevent costly architecture decision mistakes and ultimately expedite your project timeline by having a strong architecture plan.
Here are four recommendations for planning your intelligent edge architecture:
- Match your architecture to your use case
- Understand the challenges of using real-time data
- Choose an appropriate state model
- Optimize for data locality
My next post dives into each of these four recommendations, and look to identify relevant best practices and potential traps relating to each recommendation. Regardless of your use case, planning can help set you up for success. By doing your homework on the frontend, you’ll be ready to achieve your intelligent edge application objectives.
Let us know what you're building using the open source swimOS platform. You can get started with swimOS here and make sure to STAR us on GitHub. You can also learn more about our enterprise product DataFabric here.