High performance technologies can support the delivery of efficient and effective data management, but they are not the only tools for the task as financial firms strive to centralise reference data, deliver the right data at the right time, and move towards on demand, always available data. This was the assessment of a panel of experts discussing emerging technologies for the data management segment at last week’s A-Team Data Management Summit (DMS).
Moderating the panel discussion, Andrew Delaney, editor-in-chief at A-Team Group, asked panel members their views on the impact of technology on data management and how it can support user requirements.
Stuart Grant, EMEA business development manager for financial services at SAP, stepped back in time before fast forwarding: “Legacy systems implemented 10 to 20 years ago and siloed systems won’t satisfy regulations and business going forward. Considering regulation around risk there are many pieces of regulation that have a common demand for a view of risk across an organisation. The challenge is to create a plan that looks out five years and integrates reference data with distributed calculation engines, reporting and downstream analysis. The need is for a single version of the truth.”
Pavlo Paska, director and senior consultant at FalconSoft, added: “Banks want to make decisions on accurate, live data. The challenge is how to deliver consistent data at the right time to the decision maker.”
From a user standpoint, Rupert Brown, lead architect in the CTO office at UBS Investment Bank, noted the need for the right data at the right time. He said: “We are interested in latency and moving data as fast as possible, and we also have a fetish about formats. The aim is wholeness of data, how we look at it and how we deal with it.” His concern is that there are no data discovery tools of real value in the market.
Turning to emerging technologies, John Glendenning, managing director of EMEA at DataStax, provider of an open-source noSQL database, said: “There are legacy relational database management systems that firms are pushing to go faster and in-memory technologies that are used to cover up any gaps in performance. Then there are companies like Yahoo and Google investing in something new – on demand, always available big data that can be managed by systems like DataStax’ Apache Cassandra noSQL database. In financial services, the database is being deployed for risk and end-of-day pricing, allowing models to be run more frequently.”
While new technologies have a place in the financial industry, transition is not always easy. Grant explained: “Changes in technology over the past five years have been incompatible with keeping the lights on, as keeping the lights on and a strategic approach of breaking down fragmented architecture and turning it inside out don’t mix well. Covering up the cracks doesn’t solve the problem, but new technologies, expertise and relevant old technologies, such as data dictionaries and metadata, can be combined in a new approach.”
In terms of specific emerging technologies, Delaney questioned the use of cloud services. Brown responded: “The price advantage of cloud has been the driver of interest, but with it come the challenges of meeting privacy regulations and engineering the technology to make cloud solutions truly dynamic. A lack of good tools in the market makes it difficult to provision data dynamically.”
Brown went on to talk about Hadoop, an open-source software framework that supports data intensive distributed applications, and suggested Hadoop 2 tools will bring the technology further into the financial sector. He also pointed out that, perhaps converse to perception, many banks are trying to manage less data, not big data.
One common theme in the panel discussion was access to data and improved analytics. Grant commented on the need to consolidate workloads into real-time enabled datasets, a solution that is being driven by risk. In terms of analytics, he said: “Rather than trying to bring fragmented data together, it is possible to take queries to where data resides. This pushes business rules closer to data.” On the issue of access to data, Paska added: “Meta data can help, it describes data and makes it accessible, and it is a way to make sense of data and make money out of it.”
Considering data management on a longer-term basis, panel members concurred that company culture, understanding and communication at a high level are key to success. Looking at trending technologies, Brown concluded: “Semantics and ontology are the new frontiers in data. We haven’t yet joined together the systemic and semantic worlds, but we do have a grand unification plan.”