When Andy Jassy announced the release of AWS Greengrass in Limited Preview back in November at AWS re:Invent 2016, the show floor was buzzing for hours. And for good reason – AWS Greengrass promises to power up what we call “the Edge” by allowing connected devices to run local compute, messaging, data caching, and synch capabilities securely, even when they’re not connected to the Internet.
And while connectivity stands at the core of IoT, if you’re developing or deploying connected devices you know you can’t always count on a steady connection. AWS Greengrass also lets one device act as a hub for others, so they can save energy by keeping data connections with a local intranet.
So the recent announcement of general availability for AWS Greengrass is a big deal for IoT developers. It’s an important new component in the Cloud toolkit that powers IoT, and gives us 6 things to consider immediately:
Key Takeaways from AWS Greengrass Announcement
1. Software Skills Are Now More Portable to IoT
Greengrass moving more of what software developers do to the edge of the network is a major inflection point.
The enormous software developer community is already moving into IoT, and tools like AWS Greengrass will help them to dominate the space over time. Their familiarity with programming in languages like Python and writing lambdas to process streaming data in-flow now becomes applicable on the other side of the “choke point.”
2. Connectivity is More Important Than Ever
That choke point has always been, and will continue to be, cellular opex. When it comes to connectivity, IoT developers need solutions that are simple, reliable, and flexible.
Current product offerings rarely deliver on all three, and often burden projects with significant commercial and technical constraints even during development and testing. Signing complicated contracts in the prototype stage, when there’s no guarantee that devices will even work as designed, can still require a years-long upfront commitment that still amounts to small potatoes from the provider.
And if that weren’t challenging enough…
3. Efficiency Still Matters, Even at the Edge
Each meaningful communication between a device and the cloud comes with the burden of retransmitting repetitive, undifferentiated data related to security, encryption, and packet headers.
A significant portion of IoT cellular opex – from as “little” as 25% to as much as 75% or more in many cases as – is related directly to this overhead or “connectivity tax.”
It would be highly preferable from both a performance and a cost perspective to transmit only the meaningful data: which device is this, what does the register read, and nothing extraneous.
4. Your AWS Competency Now Unlocks Edge Computing Capability
While software developers are naturally curious and like to learn, sometimes it’s great to be able apply existing skills to an expanded set of problems. As offerings like AWS Greengrass bring the “software development environment” further downstream to ever-more-constrained device environments, the horsepower, elasticity, and agility of the cloud environment will provide greater and greater leverage.
More importantly, the hundreds of thousands of developers who are already capable of writing functions in lambda can now process large streams of data on the edge of the network, and transmit only the relevant information across the cellular network.
Think, for example, of a video streaming application to which a filtering function is applied, sending only the relevant information (car make, car model, number plate, time of day, GPS coordinate or space number) to the cloud when exception conditions are met.
5. The Total Package for IoT is Almost Here
Device-based environments like AWS Greengrass promise a new, easier IoT featuring:
- Familiar development tools
- Pre-processing on constrained devices
- Minimized cellular data connectivity (or newer variants like RPMA, NB-IoT, CAT-M)
- A flexible, take-what-you-need, pay-as-you-go fee structure.
So what’s missing? Security.
6. But the Fundamentals Still Apply
The expanded attack surface in IoT may be unfamiliar to software developers entering this space for the first time. It’s a reasonable source of concern, but help is on the way.
Rapidly deployed, securely managed connections to AWS that run on a VPC peering platform with the same take-what-you-need, pay-as-you-go fee structure offer a reliable solution for developers coming to IoT from the software side.
After all, if the devices you’re using to create your IoT solution aren’t publicly visible, they’re not reachable from the public internet.
Keeping in mind that the base operating system needs to be managed and regularly updated, the combination of AWS Greengrass type architecture, secure hardware elements, and an “invisible” peer connection can drastically reduce the attack surface.
For developers used to working with software, physical devices present plenty of new challenges. They’re fussy, they’re messy, and they can’t just be overwritten with a dot release.
AWS Greengrass simplifies the interface and makes software skills directly applicable to IoT. Combined with a smart approach to connectivity, it’s a huge step toward making the IoT accessible for all, and we should expect to see some exciting new use cases emerge as the developer community starts to put these new tools to use.
CJ Boguszewski holds an MBA from the London Business school and leads Americas business development for Soracom, a leading global provider of secure, scalable, cloud-native connectivity for IoT.