Big data analytics company Guavus is expanding both the breadth of data it takes in and the speed at which it can use it, with its latest Reflex 2.0 release, including support for Apache Spark and Hadoop Yarn.
Apache Spark and Hadoop Yarn are both part of an open-source analytics core. Spark is defined as the general compute engine for Hadoop data, while Yarn is the framework for job scheduling and cluster resource management. By supporting them, Guavus Inc. says its software can analyze data in the context of other events in the enterprise, even if they're in a different business silo.
Guavus says the new and improved Reflex Operational Intelligence Platform expands the ability to provide real-time analytics that pulls from multiple sources and systems, including billing, CRM, OSS, network, apps, devices, and the cloud. The expansion is important from both a real-time analytics standpoint and a continuous data correlation and actionable insight perspective, according to Heavy Reading analyst Ari Banerjee.
"In the telecom environment we talk about real-time analytics capability, which combines data existing in siloes with real-time customer transaction data to achieve results, which enables assurance, troubleshooting, personalized offer management, and so on in near real-time, taking into account as much static and dynamic context of the customer as possible," Banerjee says. "This approach is a step in the right direction."
Guavus's software is based on streaming analytics to ensure the data is actually coming in real-time, and the vendor says it can be used to solve network operations, marketing, care, and security business problems. Reflex also supports virtualized and elastic architectures, meaning it plays nice with operators' NFV and SDN strategies.
Guavus has been on the scene since 2006, before the term "big data analytics" took the telecom market by storm, and it's quickly racked up an impressive customer roster, including four of the top five mobile network operators, three of the top five Internet backbone providers, and 80% of North America's cable providers. The company claims to analyze more than half of all US mobile data traffic, processing over 2.5 petabytes per day. (See Big Data Attracts Big Dollars, New Faces, For Guavus, Next Stop AT&T?, and Amdocs Wants to Be Big in Carrier Big Data.)
As the amount of data it ingests continues to grow, Gauvus has also been working on moving closer to the radio, helping operators make sense out of RAN data. Today's announcement is another step inwards towards seeing the big picture across silos and making sense of it. (See Guavus Analytics Reverb in the RAN and ConteXtream, Guavus Forge Partnership.)
"Our recipe and model is technology, domain expertise, and data science," Guavus CMO François de Repentigny said in an interview at TM Forum Live! "Our platform embeds these things together in a package that works in a unique solution for a customer."
— Sarah Reedy, Senior Editor, Light Reading