Skkynet Industry Perspective: MQTT accomplishing things you couldn’t dream of doing 10 years ago
Here we chat with Xavier Mesrobian, vice president of sales and marketing with IIoT solution-provider Skkynet, to explore how machine-to-machine network protocol MQTT is evolving alongside our digital transformations and finding new applications in our smarter working world.
Q: What does MQTT mean to you?
A: MQTT is a great opportunity. It is a protocol that's not new, but it's the ‘old new’ for industrial automation. It's been around for many years, it's just all of a sudden become the new protocol to use or the new messaging transport to use, if you will.
It’s going to open up a lot of doors to a lot of systems integrators, and there’s going to be a lot of software and services going to be wrapped around MQTT. So for me, it’s a great opportunity.
Q: Why is it getting so much attention in the industrial space?
A: So MQTT, which is Message Q Telemetry Transport if you can remember those terms, is a published, subscribed messaging protocol that was originally created for resource-constrained devices over low-bandwidth networks. It’s being actively promoted as an IoT protocol because it’s really a small footprint. It’s reasonably simple to use. It’s battery friendly, and it features a push architecture. So MQTT works by allowing data sources like hardware devices to connect to the server (called a broker) and publish their data to it. So any device or any program that wants to receive that data would publish to that channel.
The real benefit of MQTT is its push architecture, which is inherently more secure because it avoids the client-server architecture that we see with Modbus or OPC. It allows devices to make outbound connections without opening any inbound firewall ports. And by using that central broker, it’s also possible to establish many-to-many connections, allowing multiple devices to connect to multiple clients.
Let’s put it in context—if I’m monitoring remote tanks out in the field, using MQTT is ideal for the edge devices because it never exposes the edge to an attack. That’s why it’s gaining a lot of traction, because with cybersecurity being a top issue now, people are finding ways to keep those ports closed.
Q: As edge devices become more connected, as we perform more computing at the edge, as connectivity snowballs year-over-year, how are the applications of MQTT and that evolving space changing?
A: If I look at the early implementations of MQTT, it’s always been about remote monitoring, just getting the data, connecting that data. But today we’re doing much more processing at the edge and you can see that in terms of the implementations. The value that we’re getting is significantly more. When I take MQTT and publish it to AWS, Google, Microsoft or some other broker, I’m able to run analytics, deep-learning models, predictive-maintenance models. But we’re also looking at MQTT beyond remote monitoring and into control. Its value is starting to change. And as we start talking about control, we need to have edge processing, but we still want to get that data across the pipe.
Q: What about emerging obstacles? What challenges exist with MQTT in this enhanced capacity?
A: There’s a lot of challenges, but you need to remember that MQTT is not really a protocol, it’s a message transport. So each message that is sent by MQTT could technically be in any format. Since there’s no defined protocol in MQTT, you can have an interoperability issue when you’re dealing with multiple devices. The job of the broker is not to read the message, but to relay it to other clients. So the broker is essentially dumb in a traditional MQTT environment; it can’t intelligently route or guarantee message order.
One of the other main obstacles is there’s a unique relationship between the client and the broker, which is one-to-one. So if I have a client that’s publishing and another one that’s subscribing, the subscriber has no idea of the connection status of the publisher, as it only manages its own connection and not the connection of the chain. So that creates a problem when you’re trying to daisy chain brokers or deal with quality of service.
Q: So how do users overcome these challenges with MQTT?
A: Our DataHub software was designed to resolve the inconsistencies within MQTT. We’re a smart broker. DataHub is data aware—aware of what’s happening in the system. It’s connection aware; it can intelligently route and secure data to connecting clients, regardless of the package type. Each client knows the connection status, the data is delivered in order, you can restrict the topic access, you can mark the data as read only, and it’s integrated with other protocols like OPC and Modbus. DataHub solves that edge case that a standard broker will not. But, more importantly, because we're integrated with the other protocols, now we can develop solutions, or our clients can develop solutions that bridge OPC and Modbus. And now they have a homogeneous solution that provides what they require.
Q: Okay. Let’s look at the big picture. Is MQTT the key that unlocks true capabilities with Industrial Internet of Things approach?
A: In a simple word, MQTT has to be in your toolbox for IoT. But is it the be-all end-all for IoT? I don't believe so. You’re actually going to be using multiple protocols based on the use cases or based on the actual application requirements. As an example, when I'm dealing with leak-detection systems, to do leak detection system criticality is important. Is MQTT right in that space? No. But, is MQTT right to do the diagnostic capturing of the sensor? Absolutely.
MQTT clients are just one part of the puzzle. The broker is actually the critical component to make the glue work. That's where you need to make sure that you have the right broker in your implementation.
We're seeing a lot of initiatives come out with respect to MQTT and MQTT Sparkplug B. It is growing. The device manufacturers have all jumped in. In the next few years, we're going to see more applications using MQTT than we've ever seen. And because it's such a low cost solution, we're going to start seeing sensors and big data analytics in areas that we never thought were possible. And so from our perspective, we see it as a great opportunity. I am loving the applications that we're getting involved in. They're just so cool. People are doing stuff that you couldn't dream of doing 10 years ago.
For more information, visit skkynet.com