Accedian launched a service analytics platform, DataHub IQ, to extract granular insights from the active and passive network and application performance data generated by its performance monitoring platform Skylight.
Skylight is a virtualized network services platform that provides service activation testing, network fault isolation, bandwidth monitoring, and service assurance functions. It also applies NFV monitoring capabilities at various locations including on premises, at a central location, or at the network core.
The Skylight platform combines active and passive testing capabilities, but according to Tom Foottit, the senior manager of network analytics at Accedian, the company lacked a tool that could take all the data it was creating and analyze it to provide insight and predict network issues. Accedian tried letting customers use existing analysis tools on their networks, but it wasn’t successful so it decided to build its own platform.
“The vast majority of deployed software in this space was designed and built off technology that is at least a decade old in many cases, if not more,” Foottit said. “The primary difference is they weren’t built to handle large volumes of time series data.”
The DataHubIQ augments granular Accedian performance statistics with network data from other vendors and network sources. In order to handle the large volumes of time series data, Accedian built a big data, machine learning, artificial intelligence (AI) stack into the software using open source technologies and algorithms. This AI stack enables it to perform pattern detection, anomaly detection, and data correlation in real-time.
Accedian used technologies from Apache for open source big data and time series data stock including Apache Druid, Apache Spark, and Apache Kafka.
One of the main use cases that Accedian sees this as benefiting is the automation of performance in virtual and hybrid networks for closed loop automation and self-healing networks.
London-based service provider Colt was one of the first to trial this technology. Accedian integrated with Colt’s on-demand system to build a dynamic, on-demand network service offering. The offering is a flexible service, instead of a fixed service, that enables Colt to scale up its network and adjust the service perimeters in a closed-loop system to meet increased needs automatically without any rules of human intervention, said Foottit.
Accedian also sees the platform as serving in 5G and IoT use cases. “People are going to need network slices that have a certain performance to them and being able to measure performance in milliseconds requires microsecond accuracy in your measurements,” he said. DataHub IQ can measure the performance of these backhaul, fronthaul, and mid-haul architectures to assign network slices to specific users based on the granular performance data.
Currently, DataHubIQ is launching with performance data on active monitoring metrics, but as it evolves it will combine the active and passive view together, according to Foottit. This will show the impact of network issues on applications and more.
Earlier this year, Accedian acquired Performance Vision to integrate its APM/NPM technology — primarily its end-user digital experience monitoring capabilities and wire data analytics — into the Skylight platform.