SAS, Cisco claim first platform for IoT analytics at the edge

Networking giant Cisco and business intelligence company SAS claim to have developed the first ‘edge-to-enterprise’ platform for IoT analytics.

Driven by a growing demand from customers who want to ‘analyze data in motion’ – meaning before it is stored or real time – the SAS, Cisco IoT Analytics Platform has been developed to enable analytics on devices at the edge of the network.

The point of this, as noted in a report on the product by analysts at 451 Research, is that there is little value in analyzing some kinds of IoT data “after weeks, days or even in some cases, hours.”

Out on the edge

The platform’s reference architecture combines Cisco’s networking, edge and datacenter infrastructure with SAS’s streaming and advanced analytics capabilities.

There are three main elements that end users will benefit from, say the companies. At the edge, so out on the device, Cisco’s 829 Industrial Integrated Service Routers can now run SAS Event Stream Processing (ESP), which is what enables analysis of data in motion with a ‘sub-second response time’, close to the devices and sensors generating that data.

Cisco’s 829 routers are supposedly designed for deployment in harsh environments, which could be useful for the kinds of customers the two companies say they are targeting, including those operating oil rigs and smart grids.

The second feature is flexible enterprise computing, whereby relevant data sets are identified at the edge and sent back to the data center or cloud and combined with historic enterprise data to add context. This is done using the open source publish and subscribe mechanism, Apache Kafka.

Last up is network management, connectivity, management and monitoring. Cisco will provide the robust infrastructure to connect the edge devices to the data center or cloud, and will support the management of analytics at the various network layers. Again, this is necessary for businesses like electric utilities where a large number of edge devices are used in things like smart grids.

Millions of events

Cisco and SAS said they have validated their design using sensor data from a smart data containing millions of events. According to 451’s report, the data was ingested by SAS ESP deployed on the Cisco 829 routers. It was then streamed to an ESP instance in a data center and finally into Apache Hadoop for further analysis and for models to be built.

The companies say that, as well as the aforementioned use cases in utilities and oil rigs, the platform would benefit companies manufacturing connected cars, as well as anyone looking at asset performance or predictive maintenance.

SAS admitted that its ESP does not currently support the statistical language R, but said it may add this if customers ask for it. Instead, models can be written in C++ or Python today.

Nevertheless, with market research company IDC predicting that by 2019, 45 percent of IoT-created data will be stored, processed, analyzed, and acted upon close to, or at the edge of, the network, this is an important step.

Edge-to-enterprise

Alan Webber, IDC’s research director, said the ability to turn IoT data into valuable knowledge is what “makes this partnership between SAS and Cisco so exciting. It brings two world-class companies together to offer truly edge-to-enterprise value in transforming edge data gathered through IoT into unmatched business value through analytics.”

Mike Yost, president of MESA International, a global community of manufacturers, producers, industry leaders and solution providers, also sees value in the platform. “When you dig down into the heart of the IoT discussion, it’s all about the business value potential presented by bringing IT and OT together – a complex challenge our members are eager to solve,” he said.

“Confidently pushing analytics through the network to where the data is created solves many problems. I trust people will enjoy being transformed from data caretakers to explorers, discovering all the possibilities that exist with better insight.”

451 analyst and author of the report, Jason Stamper, said “No single vendor has what we might call a ‘complete’ Internet of Things (IoT) technology stack, and SAS and Cisco are no exception. However, by working together, they have shown how it is possible to do analytics from the edge of the network right through to the data center and various types of cloud, too.

“It may be a reference architecture and a coming together of what were existing products, but for companies looking to do IoT-type analytics, it still makes a lot of sense. It also highlights the importance of edge computing for its role in IoT, as well as the leading role that stream processing technologies are likely to play in the era of the Internet of Things.”

Full analysis of the product can be found in 451’s report SAS takes Event Stream Processing to Cisco’s Edge for Internet of Things Stack.