Kathy Gibson at VMworld in Barcelona – The network isn’t what it used to be: in the digital world, the network now extends from the cloud to the core to the edge – adding to the CIO’s concerns about management and security.
Nowhere is this more apparent that at the edge, where the Internet of Things (IoT) is starting to gain real traction, says Joe Baguley, chief technology officer: EMEA at VMware.
There’s no shortage of sensors being deployed in “things” as diverse as manufacturing plant, city infrastructure and even building materials.
Meanwhile, the narrow-band networks, data collection gateway appliances and analytics to make intelligent use of the data produced are all well understood.
But there’s a fundamental piece of the IoT puzzle that’s still missing – and that’s the management of the IoT network infrastructure itself.
Sanjay Poonen, chief operating officer: customer operations of VMware, believes that artificial intelligence (AI) has a massive role to play at the edge, to manage the sensors that make up the Internet of Things (IoT)
“A huge amount of data in the future will be generated from edge devices. This means a lot of compute will have to move closer to those devices.
“This is good news for VMware – these devices will have to patched, secured and managed. And that is something we are beginning to focus on.”
One of the biggest challenges with the extended edge environment is around security, Poonen adds.
“These devices are not housed in a safe environment: they are out on the street, in elevators, on the top of traffic lights. So we will need to find new ways to protect them.”
Baguley explains that VMware is working on two separate but related technologies that will address this issue, and help IoT solution providers and customers to better manage these increasingly-vital and –pervasive networks.
LIOTA (Light Internet of Things Appliance) is open source software, developed in Python, that will run on “just about anything”, Baguley says.
“This LIOTA sits on a gateway that is positioned either inbetween sensors themselves, or between sensors and agents. It gathers and processes the data being sent from the low-energy, low-bandwidth devices in that particular IoT network.”
The second technology, Pulse IoT, sits on the backend and allows administrators to get answers to questions about the sensors. “These questions include things like: Are they alive? Are they secure? Are they patched? Can I talk to them securely?” Baguley explains.
“Pulse IoT also provides a platform for vendors to talk to the devices and sensors on the network, for instance to push firmware upgrades out to them, or to manage them in other ways.
“With these two innovations, VMware is becoming the infrastructure provider for IoT.”
The key to enabling this is NSX, which provides security across the whole network, from edge to cloud, Baguley adds.
These developments are new, and VMware is still working on its route to market. It will probably be a combination of traditional partners, IoT solution vendors and new partners, he says.
The IoT innovations have come out of the VMware research laboratories, and Baguley says there are many other developments underway.
“For instance, we have research going on into artificial intelligence (AI). Our focus is on understanding how users are consuming AI and looking at ways of supporting their AI hardware.
“It’s what we call AI for the 99% of people out there – we are talking about real-world AI applications.”
IoT is one of the areas where AI is showing a lot of promise, Baguley adds.
“IoT has turned the Internet upside-down: the Internet is traditionally about a company or individual sharing information that anyone can look at; but IoT is about hundreds or thousands of devices sending information that one person or system looks at.
“You simply have to have some level of filtering and intelligence for this to work.”
Machine learning is another area where VMware is investing, to add these capabilities to its products.
Skyline is an early example of this. It’s phone-home intelligence for customer installations, Baguley explains.
“With Skyline, agents collect data from servers in all manner of forms. This data is sent to us, usually anonymised, and this allows us to be more proactive with our support, upgrades and product development.”
Skyline is just beginning to be rolled out now, initially in the US only.