Home Availability What Does Reliability Have to do with the Internet of Things?

What Does Reliability Have to do with the Internet of Things?

by Jason Andersen

A lot of our attention these days seems to be on building highly reliable (up to carrier grade) clouds, but another area we’ve been collaborating with partners on is the notion of the Internet of Things (IoT). This is a fascinating space that, as is the case with cloud, will be transformative for IT operations. But when I speak to people about this they often look at me a little funny and say “You don’t make mobile devices. You make big servers. What do you have to do with IoT?”

Well, as it turns out, it’s a lot.

Most of the IoT attention these days is placed on the end points and the network that carries that data. And yes, that’s a big deal, but in reality there are many layers that support an IoT architecture. If you google “IoT reference architecture” you will see a lot of pretty pictures that define these layers. While definitions vary from vendor to vendor, at a high level the layers are:

  1. Devices (aka the Things) – these are the specific endpoints in the operations and could be anything from a driverless car, to a security dongle on a palette in a warehouse, or your Apple Watch.
  2. Access Network – This is the network layer that the device uses to get to its first access point.
  3. Control Layer – This is probably the most loosely defined layer since it can encompass a lot of technologies. It can be deployed at the edge, in the data center or in the cloud and it can support a wide range of applications.
  4. Back End Services – This is where the action likely happens for data analytics and aggregation, chargebacks or other centralized services

Now it’s safe to assume that the Back End Services layer is likely to be managed by the IT team in such a way that it’s reliable and available for the rest of the layers. However, the Control Layer can be an entirely different matter and you really need to think about reliability at that layer. The reason reliability is so important may not be as straightforward as you think, but in IOT, the real action is in the Control Layer. Traditional application architectures rely heavily on back end services to deliver real time support to the end systems (usually people driven devices like a laptop or tablet). However, in the IoT world there is a lot more real time activity in the Control Layer.

Applications sitting in the Control Layer include:

  1. Process Control Applications – All of these devices themselves will have some degree of automation and the actual delivery of that automation will be planned, communicated and possibly monitored at the control layer.
  2. Compliance Applications – Some IoT environments will be either regulated or audited at some point. The devices will pass data to the Control Layer which may or may not be further shipped up to the Back End Services for analysis. However, we still often see the need for a more localized compliance solution for network or other business reasons.
  3. Messaging Applications – Even if you are not storing the data or managing the devices locally you may still need some sort of security or messaging gateway to connect the edge to the core.

So, the Control Layer is really the unsung hero of IoT. In the data center this may be simply lumped in with the rest of the Back End services. There are however many cases where the Control Layer sits outside of the traditional data center (what we at Stratus call the Edge). At the Edge, you need more than just reliability. You also need simplicity due to a lack of IT skills at the Edge. You will also need longevity, since those machines tend to stay in-situ for way longer than data center systems.

Related Posts