We’re living in a centralized world – but that’s all about to change with the proliferation of decentralized, intelligent IoT devices, sensors and machine to machine (M2M) communications. Big data, security, governance and compliance are just a few of the key challenges businesses face when programming applications for the IoT landscape.
This is the road ahead, according to enterprise cloud company Nutanix. The company’s vice president of technology Satyam Vaghani has detailed the Nutanix IoT vision at its .NEXT conference, taking place this week in Washington DC.
New paradigm, new OS
“Every time there is a fundamental change in IT paradigm, a new operating system is needed,” said Vaghani. “Look at the progression from desktop to mobile. Now, we also see the same thing happening in datacenters feeding the IoT from the cloud.”
Vaghani suggests that, in the new model, cloud no longer works as the central ‘brain’. More and more of our computing will move out to the edge (both the edge of the device and the edge of the network). The key change, he claims, will see the edge ultimately trump the cloud in terms of where things happen.
“The edge trumps the cloud.” You can almost see the T-shirt designs already.
This shift is already happening and there are already some observable trends to validate it.
Terabytes in aircraft IoT engines
Look, for example, at the amount of sensor data being created on modern commercial aircraft today: according to Vaghani, two General Electric aircraft engines with 5,000 sensors mounted on each will generate a collective 1 terabyte of operational data every day.
“There is never going to be enough physical time to transport all of that IoT data to the cloud,” he said. “What we will now see is the IoT processing real-time data at the edge and the cloud working at another level to process what we might call long-term data.”
He added that the hard part here is that some elements of a single application will have to run on the edge, while others will have to run in the cloud. This creates infrastructure problems when applications out on the edge need to be patched or updated.
Things get even more complex from there. The entire system has to be multi-tenant because, as sensors get more powerful, different employees will want to access them.
Vaghani offered the example of a security camera system. For obvious reasons, it can’t be accessible to everyone in an organization, “so being able to segment differentiated data streams of information, based upon different employee policy access levels, has to be architected.”
Unification, to balance decentralization
This explains a large part of why Nutanix is aiming to bring a more unified approach to software-defined, infrastructure-centric computing with its latest product updates. The company insists that a single software operating system brings a new approach to unifying multiple clouds, across the full compute, storage and network ‘stack’.
Nutanix is looking to simplify operations with common IT tooling, enabling application mobility across clouds, while remaining open to any hardware, hypervisor, or cloud.
The Nutanix Enterprise Cloud OS will be delivered as a ‘full software stack’, with new multi-cloud capabilities in Nutanix Calm (a DevOps, abstraction and software operations control technology) and a new cloud service, called Nutanix Xi Cloud Services.
Bringing all this technology into functional deployment necessitates the abstraction of software-defined system building blocks. If we make these blocks of functionality definable in this way, then we can get whatever we want out of the data source in question.
Looking at news announcements from Nutanix in the IoT space this month, we can see that the company has agreed to form a new alliance with Google.
This focuses on Internet of Things (IoT) use cases that marry real-time edge intelligence with core cloud computing. Customers can use Nutanix as an ‘intelligent edge’ for IoT applications based on Google Cloud Platform (GCP), by deploying TensorFlow for edge processing, while training machine learning models and running analytics on the processed metadata in GCP.
A new road to a new universe
“This new universe of machines begs for a new type of operating system, one that spans both edge devices and the cloud to offer the seamless layer of services demanded by most IoT applications,” according to Vaghani, writing in a company magazine.
The new road ahead to IoT systems and applications is typified by new network traffic patterns – and these new patterns require inherently different new design and capacity planning.
Agreeing that the immediate road ahead presents challenges in terms of IoT platforms and architectural refactoring, Vaghani and team point to a brighter future where standardized security and the embrace of a more consistent operating system for the IoT makes things simpler.
The IoT has grown up fast and, for many, the initial understanding was that the IoT runs on the cloud. Now, they are finding out (if they buy into the Nutanix message) that the IoT runs on the edge. Yes, edge computing is massively dependent upon and connected to the core cloud model. After all, it evolved as an extension of it – but now, it seems, it’s possible to go several steps beyond it.