Research Results

The researchers Zsolt István (IMDEA Software Institute), Kaan Kara, (Oracle Labs), and David Sidler (Microsoft Corporation) have published the book “FPGA-Accelerated Analytics: From Single Nodes to Clusters”.

Today datacenters that host data-intensive applications used in online services and machine learning need to store and process data that increases at an exponential rate. Data processing and management applications have become increasingly distributed and this has lead to new data movement bottlenecks at various levels of software and hardware architecture.

The authors survey recent research on using reconfigurable hardware accelerators, namely, Field Programmable Gate Arrays (FPGAs), to accelerate analytical processing. Such accelerators are being adopted as a way of overcoming the recent stagnation in CPU performance because they can implement algorithms differently from traditional CPUs, breaking traditional trade-offs.

Zsolt, Kaan and David discuss the benefits of using FPGAs in the context of analytical processing, both as an accelerator within a single node database and as part of distributed data analytics pipelines. They also present guidelines for accelerator design in both scenarios, as well as, examples of integration within full-fledged Relational Databases.

Finally, they highlight future research challenges in programmability and integration, and cover architectural trends that are propelling the rapid adoption of accelerators in datacenters and the cloud.