The intelligent edge is still a new concept for many potential customers. So what is it and why do they need it?
Gareth James, network and security sales specialist, VMware: The intelligent edge is important for three reasons. You need real-time decision-making at the edge, where speed is important and you can't afford the latency of roundtrips. Then you've got cost reduction - the more processing you can do at the edge, the less data you have to send back to head office, so you can reduce your data bandwidth costs. Lastly, it's about business continuity, because the last thing you want is that if you lose connectivity to the edge, the business can't function.
Rudeon Snell, lead for SAP Leonardo Intelligent Technologies, EMEA South at SAP Africa: We see this convergence happening right now. Sensor prices dropped by 90% in the last five years and there's been an explosion of data use. Customers can't wait for latency in the roundtrips to process the amount of data that's being generated. If you think about smart city constructs, intelligent buildings, agri-solutions, areas where cloud isn't pervasive, you have to have alternate solutions in place that can simulate the experiences that customers are after. It's a relatively new technology that's projected to go mainstream in the next five years only.
George Senzere, solutions engineering manager, Schneider Electric: It's defined in lots of ways, but I like to see it as a place where the real action is. You're talking about smart buildings, smart factories, your house, it's everything. What's driving it is the deluge of data generated by the IoT. You have to look at reduction and deduplication of data. Then there's protection of data, and where data is stored. Intelligent edge can help with compliance around data sovereignty laws, too.
Rudeon Snell, SAP: What I love about the notion of the intelligent edge is that it's the glue you play with to aggregate data from multiple sources into a cohesive layer from which you can make even more informed business decisions. It's not just about dealing with forward-facing integration of IoT, but also integrating data from legacy equipment and taking all of that data into a layer of abstraction in the cloud, from which more informed business decisions can be made.
That's why intelligent edge is being built here; African infrastructure isn't reliable.
Tobie Alberts, IoT.nxt
Tobie Alberts, digital evangelist, IoT.nxt: Operational Technology (OT) has been around for a long time. We've been doing intelligence at the edge for a long time. Why is it important now? Because we're marrying OT with IT, and we have the ability to have IT run on the edge, side-by-side with my systems. When we talk about IT, this is nothing new - we've been talking about the client/server relationship forever and it goes around in a circle over time; first, everything happened on the server, then it all happened on the client, then it all happens in the cloud again. What is happening is that people who manage businesses are now looking at OT and saying, 'I can manage my servers from my phone, so why can't I manage my printing machine or infrastructure from my phone?' To do that, you need intelligence at the edge.
Paul Williams, country manager, Fortinet: WiFi and 5G will be the glue that holds the Industrial IoT (IIoT) together. The mobile operators are planning on offering 5G; they're not there yet because it's going to need the buildout of infrastructure and they don't have licences yet.
Tobie Alberts, IoT.nxt: That's why intelligent edge is being built here; African infrastructure isn't reliable. It's not going to happen in the US, where they have fantastic cloud solutions or in Europe, where there's great narrowband for a few euros a year.
Rudeon Snell, SAP: We understand the value of the solutions that we propose, but all of this rises and falls on customer adoption and consumption of intelligent edge solutions. That will only happen when customers see the value of those solutions. The metrics that matter to that organisation must be positively impacted by what you propose. They don't care about the tech, they care about how it affects the top or bottom line. A lot of customers still feel that intelligent edge is a nice-to-have.
So is intelligent edge a confusing sell, because customers are still only just adapting to cloud?
Riaan Graham, regional director, Ruckus: The way the industry used to sell, those days are gone. We used to be able to just drop a piece of tin. If you can't add value to the end-customer's business, you won't get the numbers. You need to understand their core business, and, more importantly, you need to understand where the core pain points are; you'll see it before they do. They're too close to the problem. It's a consultancy process - you identify what the problem is, where the shortcomings are, and what the KPIs are.
Rudeon Snell, SAP: It's also about translation of value. If you can't land your message with your audience, and they don't get it, the sales cycle isn't going to move. How do we translate the value proposition into something that's meaningful?
Trent Odgers, cloud and hosting manager, Veeam Software: I don't like to think of cloud as a destination or of edge as a destination. It's 'and', not 'or'. As much as you're going to compute at the edge - and that brings benefit - it's still going to send information back so that you can analyse and get reports and get information. Amazon has announced Edge. You can now run Azure IoT on vSphere, the vendors are trying to work together to solve these problems. Everyone's trying to figure out how to do it with what they have, so there's some retrofitting.
Rudeon Snell, SAP: You also need political will inside an organisation to embrace this progress. It's not enough anymore to show a pain point and a solution that will demonstrate value. Those days are ending because now everyone is highlighting pains - so how do you differentiate yourself?
Blockchain is a classic example of this differentiation challenge - everyone is playing around with it, experimenting with it, so what? There's a similar challenge with edge.
Hein Badenhorst, client technical advisor and data lead, IBM: It's not just about political will, it's about trust at all levels. All the things that have happened with digital transformation - it's been oversold, a lot of people have played with it. We talk about putting something intelligent on the edge that's going to perform some business function on the business' behalf, but at the moment, many of the big CIOs find it difficult enough to control just what's inside a datacentre. Now I have to convince them that I'm going to put something intelligent out there that will operate on my behalf? The trust level has to go up exponentially.
What kinds of challenges do we see intelligent edge solving?
Riaan Graham, Ruckus: Take the big metros. We understand the challenges that all the metros have. They need to be able to do service delivery better and a pragmatic IoT strategy could do that. But they work in silos. There's no political will to share data between departments. If you look at smart meters that have been rolled out - they all have WiFi or 3G. Why don't we use that technology to start taking readings more accurately and have a dynamic environment, to fix things proactively?
Paul Williams, Fortinet: Think about water meters. People complain about paying their bills because they're too high, sometimes thousands of rands a month. If you have intelligence at the edge, you can have an active portal that lets you monitor and add a cut-off level to stop you going over a limit for water use.
George Senzere, Schneider Electric: That's the point, we understand it and we know what the value is. But for someone in a business to figure out a return on investment, it's difficult.
Trent Odgers, Veeam: Shadow IT will figure it out. Either government will get involved or businesses will get involved or users will take a couple of Raspberry Pis and just do it for themselves.
Gareth James, VMware: And that's why we can't sit and wait for 5G to happen. There's a plethora of other technologies and solution sets that will deliver results today and deliver results immediately.
Presumably for IIoT, we're going to be talking WirelessHART, Zigbee and industrial WiFi anyway?
Hein Badenhorst, IBM: One of the problems is that I spend a lot of my time talking to clients about analytics and data and why it's a good thing to do. But as soon as you start talking about the core discipline behind it, that you'll have a lot more objects generating data, and you need somewhere to store that data and someone to take ownership of it, and data scientists to look at it, and build up an API so that when something happens on the edge, it triggers a business process... We talk about it as if it's plain and simple. But the IT of that organisation can't even manage the patches on their operating systems correctly.
We spend a lot of time trying to pre-package the analytics and take the heavy lifting away from them.
It's still an adversarial system where clients are trying to squeeze extra value, you're trying to get a bigger-size deal. How do we grow the pie rather than look for a bigger slice?
Rudeon Snell, SAP
Rudeon Snell, SAP: We need more of an open ecosystem where we collaborate with customers specifically. The analogy is that we all have bank accounts, and we trust the bank to hold our money as opposed to storing it under the bed because a bank is geared up to be better at storing money. It's the same thing - our clients are struggling to deal with patches, yet they're heading off and starting data science competencies. It's about trust, it has to be a partnership ecosystem so that people can come to us and be honest and say, 'I don't know what I'm doing. This is what I want, how do we both win by building it?' Those conversations aren't happening enough, it's still an adversarial system where clients are trying to squeeze extra value, you're trying to get a bigger-size deal. How do we grow the pie rather than look for a bigger slice?
Riaan Graham, Ruckus: That's where the ecosystem partnerships become very important.
Rudeon Snell, SAP: And yet all of the vendors are developing edge capabilities that don't necessarily conform to the same standards.
Riaan Graham, Ruckus: In the IoT world now, we have ZigBee, LoRa, maybe 40 other standards and the race is on. And it should be; we haven't found the VHS from the Betamax yet. Let the reigns run free, but early adopters need to be brave enough to adopt one and be flexible.
Share