Application delivery company Apcela made its first move into machine learning (ML) with the launch of its Enhanced Analytics platform with a software-defined wide area networking (SD-WAN) API. The API leverages and extends capabilities of Cisco’s SD-WAN to increase visibility and give enterprises predictive and automated analysis into their networks, according to the company.
This builds upon Apcela’s core product and SD-WAN technology by extending data and visualization past the WAN overlay, which it was previously limited to, and takes the data outside the SD-WAN core orchestration platform.
This broadens the data set and analysis so that customers can integrate with other data sets from the application delivery stack, including switches, routers, firewalls, WAN accelerators and other network/application management platforms.
Apcela’s platform gives a more complete view of the underlay and overlay network data, compared to using only Cisco’s Tetration monitoring platform, and it isolates the problem, said Mark Casey, Apcela CEO.
“Using the Tetration Open API, data from this platform can be integrated with other data sets on EAP [enhanced analytics platform],” he explained. “For example, Tetration can isolate network performance problems via hop-by-hop views that enable you to quickly determine if the bottleneck is on the network or on the server. It doesn’t necessarily tell you whether the problem is related to the underlying carrier transport, your switches, routers, or exactly what the problem is. ”
The Apcela platform isolates the actual problem, he added, and it can conduct this correlative analysis between Cisco SD-WAN operational data as well as other network systems, such as Palo Alto, AppDynamics, Amazon Web Services (AWS), and VMware.
“The more systems we interconnect, however, when an application performance issue arises, the harder it becomes to correlate with confidence between the symptoms and behaviors observed in one system with the root cause that may often exists in another,” Casey said.
By integrating the network data from the application delivery stack and the application environment it makes it easier to identify the correlations and root causes of issues like latency, packet loss, and jitter across these disparate systems.
Historical network data allows the platform to back-test correlations and algorithms that drive automated and intent-based actions, which will allow it to identify issues and correlations in the future.
The data integration component of the platform is important, Casey said, as Cisco’s vManage orchestration platform will only retain six months of data. By moving this data to Apcela’s enhanced analytics platform, it can retain a full data set, “or whatever a client determines is economical, based on storage cost considerations,” he said.
Casey also noted that performance problems on a network often stem from long-term issues, which can make them more difficult to discover. As more systems are integrated the ability to correlate all this data also becomes more difficult. Automated remediation can help address these issues.
This move into machine learning for Apcela builds on the company’s enterprise push as it furthers product development, operational improvement, and expanding its SD-WAN vendor, which SDxCentral discussed with Jack Dziak, Apcela president and COO, earlier this year.
“Ultimately AI [artificial intelligence] and ML will combine to drive a fully automated, intent-based network and application delivery infrastructure that evolves in real-time in response to changes in the networking, security, and application environments,“ said Casey.
Leave a Reply
You must be logged in to post a comment.