Because I am traveling this week and apparently need to make it to the airport on time, please enjoy this issue (and the next issue) free of commentary on the links below. But trust that they’re interesting, or I wouldn’t have included them ;-)
The AWS Greengrass service hit general availability on Wednesday
, after being announced at re:Invent in December. Greengrass, if you’re unfamiliar, is a platform for building networks of edge devices that can perform certain compute and data tasks locally, while communicating with cloud infrastructure for long-term storage, compute-intensive processing and other things you’d expect to happen on larger, more stable machines.
The most obvious use case for a service like Greengrass (or, I should note, Microsoft’s currently-in-preview-mode Azure IoT Edge)
is the Internet of Things. If you have lots of connected devices out in the field that need to process data with low latency, possibly in an area with poor connectivity, you can’t really rely on sending everything to and from the cloud. Instead, you deploy a “core” machine (it can be very small) on the edge and let all those devices communicate via their own network for anything they can do locally.
However, an under-appreciated aspect of Greengrass is that devices run Lambda functions—as in the “serverless” computing service also available on the AWS cloud—and also don’t need to be small IoT devices. Amazon CTO Werner Vogels notes in his own blog post on Greengrass
, “With AWS Greengrass, we can begin to extend AWS into customer systems—from small devices to racks of servers.” Just think about a horde of Docker containers running on local hardware and executing Lambda functions, sending some data to local storage (for regulatory reasons for example) and the rest to the cloud.
If you were wondering how AWS might bridge the gap between data centers and its cloud, or how it might spur the world toward the serverless computing architectures it’s pushing with Lamdba, I would say that Greengrass looks like a good start toward doing both. There’s still a lot we don’t know about how edge computing will play out and who’ll ultimately own the value these new architectures, but I think we can rest assured
that the cloud providers are going to get their piece of that pie.
In other news, here are 4 items from Wednesday definitely worth checking out: