Megh Computing provides a platform for accelerating Real Time Analytics using Spark Streaming and other frameworks. The solution enables seamless acceleration of applications that process streams with Machine Learning and Deep Learning algorithms to extract value from data as it is moving.

The Megh solution supports both in-line processing of streaming data and offloading of Machine Learning and Deep Learning libraries with FPGAs.

There are two key components:

  1. Hardware platform: Intel PACs (Programmable Acceleration Cards) with Intel Arria10 and Stratix 10 FPGAs and Intel Acceleration Stack running on Dell R740 servers.
  2. Software platform: Spark Streaming framework with BigDL libraries and Megh libraries for Arka Runtime and Sira AFUs (Accelerator Function Units).

The key benefits of this solution include:

  • Applications run unmodified using standard or custom APIs.
  • Arka Runtime exposes the Accelerator Functions-as-a-Service. It manages the FPGAs and supports SW fallback for the AFUs.
  • Sira AFUs deliver the actual acceleration and are implemented as libraries that get downloaded to the FPGA.

Watch Megh Computing Founder and CEO Prabhat K. Gupta describe our product architecture.

Play Video
Share this page
Share on facebook
Share on google
Share on twitter
Share on linkedin

Retail use-case scenarios

Learn more

Get the whitepaper, request a demo below, or contact us.

Request a demo
Find out how to accelerate your Real Time Analytics workloads with Spark Streaming framework and BigDL libraries in the Cloud using FPGA accelerators