Big Data for Finance

Greenplum Debuts Analytics Workbench

The Greenplum Analytics Workbench - a 1,000-node cluster that will act as a lab environment for accelerating the pace of big data innovation - is now live.  EMC's Greenplum unit will use the Analytics Workbench to test the limits of scale-out infrastructure technology, and also to re-define the models for applying big data analytics.

Bloomberg Acquires PolarLake For EDM

Bloomberg has acquired Dublin-based software provider PolarLake, and is launching a new enterprise data management (EDM) service to help companies acquire, manage and distribute data across their organisations.

Learning From Mistakes: Applying Big Data to Realtime Risk Analysis

by Daniel Wakefield, IMGROUP
www.imgroup.com

As the financial sector once again reels under yet more allegations of huge losses (this time at JPMorgan Chase) there is a more than mild sense of déjà vu here at IMGROUP.  Yet again, questions are being asked about how one desk, perhaps even one trader, can lose such a significant amount of money without anybody appearing to notice.  The jury is still out as to whether we are talking about illegal activity and losses being deliberately hidden or whether it is a case of ignorance caused by the complexity of the system.  In some ways however this is irrelevant.  The fact remains that huge losses still seem to be easy to hide – wilfully or not.

Solace Upgrades Messaging For Big Data

Solace Systems is upgrading its messaging appliances to better support high-capacity application infrastructure and "the most demanding" distributed computing architectures.  The latest rev more than doubles the data throughput of existing models.

Also new is added support for disaster recovery with the ability to automatically replicate per-topic messages and state to secondary data centres.  And new in-memory cascading caching helps accelerate distributed applications such as trade order processing, cloud computing and big data collection, distribution or analytics.  This allows caches in one location to send on-demand queries to upstream caches, when applications need data that is not cached locally.

GigaSpaces Points XAP9.0 at Big Data

GigaSpaces Technologies has released XAP9.0 - its application platform - focusing it on real-time big data applications.  Says Adi Paz, GigaSpaces' EVP of marketing and business development: "What is common to all real-time analytics systems is the need for streaming data processing, alongside batch processing, such as Hadoop provides.  However, building a system for big data apps is complicated – it must maintain reliability, scalability, and performance even at immensely large scales of data.  XAP 9.0 was specifically designed to resolve these challenges."

Numerix Taps Azure Cloud For Compute, Big Data

Risk and valuations vendor Numerix has enabled its Numerix Portfolio software to make use of Microsoft's Windows Azure, offering the the ability to burst complex pricing and risk calculation processes to public cloud infrastructure.  As well as providing access to increased compute resources, it also allows the processing of very large data sets to solve the quantitative, computational and technological challenges associated with calculating advanced risk measures.

"The need to establish consistent pricing and risk calculations daily, in real-time, has fueled huge growth in compute power demand; but the reality is traditional on-premise deployments alone are no longer a timely or cost effective  answer," says Denny Yu, product manager of risk for Numerix. "However, with a burst-to-cloud scenario, clients now have a fast, easy and low cost way to harness the massive compute power and big data resources needed to address real-time trading and risk."

What Is: In-Memory

In-memory computing is very much the topic of the moment.  For example, Gartner has just cited it - along with big data, cloud computing and extreme low energy servers - as one of 10 disruptive technologies for business information management.

Big Data Gets Focus at A-Team's Data Management Summit

Big data will be a central topic of debate at next week's Data Management Summit in London.  As well as being included in a morning 'Big Picture' panel on emerging technologies, all three of the afternoon 'Drill Down' workshops are focused on big data subjects.  The Summit takes place on Tuesday, May 22nd at the America Square Conference Centre.  Attendance is free for members of this community - but you must register online in the next 48 hours.  Register here.

Panopticon Now Supports Visualisation for Java

Panopticon has released a Java Server version of its EX data visualisation suite - incorporating Treemaps, Heatmaps and Scatter Plot displays, optimised for analysis of current and historic time series data.  Previously, EX ran only in Microsoft .Net environments.

Mellanox, AMAX, STEC Join for Hadoop Appliance

Networking specialist Mellanox Technologies has joined forces with server/storage company Amax and solid state storage vendor Stec in the introduction of the PHAT-DATA40G, an appliance geared to running Hadoop.

The appliance includes an Amax server, hosting Intel E5-2600 chips (often know as Sandy Bridge), Mellanox's 40gE networking and 960 GB SSDs from Stec.  Installation and management tools from Zettaset are also part of the offering.  The companies are claiming a minimum 20% performance boost over existing solutions, suggesting sorting 1 TB of data can be accomplished in 27 minutes, using five appliances.