The leading knowledge platform for the financial technology industry
The leading knowledge platform for the financial technology industry

A-Team Insight Blogs

ScaleOut Pushes Hadoop Towards Low-Latency for Real-Time Analytics

OK, so the headline is a tad extreme, but bear with me. Recent developments combining in-memory technologies and Hadoop/MapReduce from ScaleOut Software point to a future where big data analytics and real-time processing, as it’s defined in the financial markets, could meet.

ScaleOut has just released its ScaleOut hServer V2, an in-memory data grid, which it claims can boost Hadoop performance by 20x, and can make it suitable for processing ‘live data’ to deliver ‘rea-ltime analytics’.

“To minimise execution time, ScaleOut hServer employs numerous optimisations to minimise data motion during the execution of MapReduce applications, and it can automatically cache HDFS data sets within the IMDG (a feature introduced with ScaleOut hServer V1). In addition, ScaleOut hServer’s memory capacity and throughput can be scaled by adding servers to the IMDG’s cluster. The product automatically rebalances the data set and execution workload when servers are added or removed,” says the company in a statement.

As well as boosting performance of a Hadoop deployment, hServer also incorporates Map/Reduce logic so that a Hadoop distribution is not actually required – though the company suggests its offering is not a direct replacement for Hadoop.

Nevertheless, “ScaleOut hServer is designed to be compatible with most Java-based Hadoop Map/Reduce applications developed for the standard Hadoop distributions, requiring only a one-line code change to execute applications using ScaleOut hServer.”

The big picture here is that ScaleOut – as well as other companies pushing in-memory technology – is recognising that the batch-oriented nature of Hadoop has limitations for real-time applications, such as those found in the financial markets.

While ScaleOut is today looking to boost Hadoop performance to make applications that used to take hours and minutes to execute run now in minutes and seconds, the performance trajectory could well follow that of the low-latency space, where milliseconds gave way to microseconds, and now nanoseconds.

The deployment of multi-core and multi-socket servers, GPU technologies and advances in memory will all benefit data grid vendors like ScaleOut, as well as Hadoop and other big data analytics offerings.

Related content

WEBINAR

Upcoming Webinar: Opportunities of new approaches to electronic trading

Date: 9 December 2021 Time: 10:00am ET / 3:00pm London / 4:00pm CET Duration: 50 minutes Challenged by legacy systems, less than ideal workflows and high costs, front-office trading teams lack the ability to adapt to clients’ evolving needs around integration, speed and multi-asset capabilities. They are also challenged by a capital markets environment characterised...

BLOG

Streams: The Way to Better Liquidity in the FX Markets

By Phil Weisberg, EVP of Strategic Planning and Partnerships, oneZero. Sourcing liquidity in foreign exchange trading is a herculean task for banks, one not helped by an ever-evolving thicket of global regulation and the complicated nature of FX infrastructure and trading. The traditional way of looking at how to provide the necessary liquidity has been...

EVENT

RegTech Summit London

Now in its 6th year, the RegTech Summit in London explores how the European financial services industry can leverage technology to drive innovation, cut costs and support regulatory change.

GUIDE

Entity Data Management Handbook

Following on from the success of our Regulatory Data Handbook, A-Team Group is pleased to introduce its new Entity Data Management Handbook which is available for free download. This Handbook is the ultimate guide to all things entity data: Why Entity Data is important A full review of Legal Entity Identifiers (LEIs) Where they came...