TECHNOLOGY

Open Source; Innovation in the Financial Sector

By Michael Down; Senior Solutions Architect, Elastic

The financial industry’s love affair with open source technology now runs far deeper than any simple attempts to reduce software licensing costs.

The free starting point offered by open source licenses can never be overlooked in driving initial technology adoption. For financial institutions whose competitive advantage is often built on codifying ideas via bespoke software development, this zero-cost point of entry, combined with an open, documented code base delivers a compelling argument for open source software adoption.

However, the industry’s ongoing thirst for open technology has less to do with purchase price and everything to do with reducing the management of hundreds of different, disparate systems running multiple parts of the company. Financial organisations of all types are characterised by thousands of point solutions, serving many legacy systems. Banks hire tens of thousands of offshore staff to manage each solution.

The real draw of open source in modern finance is the ability to explore and innovate with new technologies; to easily scale the solutions that deliver real competitive advantage; and to reduce the overall cost of managing vast IT infrastructures through the use of common, best-of-breed technologies.

Increased Quality with Reduced Management Burden

From early Java-infused engineering to today’s polyglot, containerised, data-driven landscape, the simple quality and technical strengths of open source have been a key factor in its adoption. In an industry which attracts the world’s best graduates based on their higher-level maths, computer science and engineering skills, open source tools offer the fastest, highest quality way to turn ideas into secure, operational systems. Consequently, open source adoption has often been driven by the developers themselves.

From open languages like Java, Node.js and Python; through common tools and the familiar methods used in GitHub, Puppet, Chef, Apache Maven and Jenkins; to common architectures built on Linux, virtualisation, or Docker and Kubernetes: these commonalities dramatically reduce the cost and complexity of hiring, training and transferring staff between projects. For companies running thousands of applications simultaneously, this factor alone will outweigh licensing costs by an order of magnitude.

The crucial thing for all financial companies is to reduce total cost of ownership. This is about being able to reuse one technology or platform across multiple use cases in different areas of the business. As discussed later in this piece, the reason why open source search is so prolific within Goldman Sachs, is that teams can use and reuse the same platform across a number of different projects. One tool set – one support team – one knowledge base; but varied projects.

Goldman Sachs knows that by being sagacious in its choice of technology, teams can use it across a gamut of problems, in a host of business areas, without needing to go back to the technical drawing board each time.

Goldman Sachs, and a host of others, are finding that working across problems on one open platform, they need far fewer people, focused on managing one underlying technology; rather than 40,000 offshore employees running individual toolsets.

One Platform, Multiple Uses: Goldman Sachs

Goldman Sachs provides a great illustration of a financial institution using one platform to solve multiple business issues. Having implemented a popular open source platform, they show how versatile search can be as a broad operational tool, incorporated into many different applications.

Elasticsearch is deployed across a number of divisions and a variety of use cases – trade tracking, anomaly and error detection, and the analysis of contract applications, to name a few. In the latter case manual analysis of paper documents would require multiple legal teams do the heavy analytical lifting. Apache Tika was used by Goldman Sachs to digitise and index contracts. Elasticsearch is then used to review each document, alerting company lawyers if the right references or terminology are not found.

Trade tracking is also crucial, giving teams a live view of the status on a trade at any given time, and also helping to eliminate both human and coding errors. Goldman’s trade tracker, gathers data from a number of sources, integrates and unifies it in Elasticsearch and then extracts useful data automatically. This reduces the need for large teams to extract the data manually.

‘Fat finger’ trading errors cost the industry billions every year, and there are plenty of high profile examples. In 2015 one of Deutsche Bank’s hedge fund clients received $6 billion in error. Anomaly detection, therefore, is a priority for Goldman Sachs. And giving its community of developers access to a search-based data management system has been key to more timely elimination of human error. On the bug detection side, around 700 developers now utilise Elasticsearch to trawl large code libraries to find all instances of coding anomalies. Goldman Sachs’ developers use Kibana to build status dashboards which alert teams to code alterations, compares code version variations and pulls together disparate data sources in a simple interface.

Dion Global: Wealth Management

Dion Global designs, develops and supports integrated software solutions for financial institutions across the globe. The company has recently won several awards for its data analytics solution, built on open source search technology.

Financial services firms have always worked to derive new insights from their data in order to improve decision making. What has changed, is how the effective use of data is increasingly becoming a core source of competitive advantage that separates the leaders from the laggards. Firms are increasingly looking to leverage business data to drive efficiency, grow revenues and improve margins. Dion Global’s Invantage Analytics module, built on the open source Elastic Stack, creates, maintains and secures large volumes of data. The module delivers financial analysis using Logstash to ingest data, Elasticsearch for indexing, and Kibana to visualise results. In production, Dion has then licensed Elastic’s additional X-Pack features to deliver additional security for the highly sensitive data and to allow teams to access inherent machine learning capabilities.

The analytics module helps financial companies derive insight from data, which can be drawn from any number of external data feeds. It helps firms to influence the way they interact with customers, competitors, regulators, the market; even shareholders. Taking high level stock trade information, gathered at various stages of the trading process, data is aggregated to provide Dion Global’s clients with an evaluation of their own customers’ stock trades. Dion’s clients can then see all daily trades and retrospective trades made over the past 5 years.

Elastic alerting shows teams when a certain trade crosses a certain threshold – spotting anomalies faster. Later, anomaly detection will be handled by the machine learning capabilities inherent in X-Pack; reducing the need for manual queries which further helps Dion Global to model their data automatically.

Swiss Life: 360 Degree Customer Visibility

Swiss Life is a major player in insurance and wealth management. A few years ago, Swiss Life France embarked on a company-wide strategic project – Digital Foundation – to re-architect and digitize its system architecture across all of its web and mobile-enabled portals and applications. With over ten million customer records accessed with divergent means and formats, it was becoming increasingly difficult to make data consistently accessible to different audiences including private and business clients, sales people, insurance brokers, and customer service representatives.

To streamline operations and unify access to information, Swiss Life France knew it needed a way to query customer records, at scale and speed, in real time. Initially, the company started working with Elasticsearch to index and publish its customer data. Mapping the system to the current and future requirements of digitalization. The company had access to real-time queries across customer records, contract data, market segmentation data, and pension and insurance scoring information.

Using Elasticsearch, Swiss Life could merge multiple heterogeneous data sets together and derive cross-source insights continuously and in real-time. Elasticsearch gathers all the customer data in one place and acts as the single point of exposure for all customers through the MySwissLife customer website and mobile application. For the portal’s users, it is critical that the system is able to look up customer and contract information in milliseconds. Equally, Swiss Life must guarantee that the information is propagated to the index in less than 10 seconds after any record has been updated in the source systems.

Innovating within New Regulatory Frameworks

Regulatory complexity has slowed the pace of innovation within banks in recent years. At the heart of this has been an inability to give teams a common platform to innovate and consume software toolsets at will. Competition now comes from technology companies, like Google, Facebook, Amazon and Apple, built around an ethos of ‘democratisation’; who provide ready access to tools across the organisation to promote innovation.

For finance companies, a large part of the attraction of open source technologies comes from their low cost, low risk starting point. It costs nothing to download and start playing with open technologies, companies have nothing to lose in making toolsets widely available within development and engineering teams. Open source platforms tend to work the same across problems – same interface, same tooling, no complex integration – so it’s easy to experiment. A new way open source technologies are enabling innovation is around new regulatory frameworks – particularly the Payment Services Directive 2 (PSD2).

PSD2 and Erste Group Bank: Transparency for Consumers

Giving consumers access to the full gamut of personal data held by their banks has become a priority as PSD2 comes into focus. Under this legislation, financial institutions must give open access to their payments infrastructure and customer data. This is to allow the industry to create transparent payments and information systems for consumption by the end-customer. EU member states must implement the legislation by 13th January 2018, with firms expected to show compliance soon after.

Erste Group Bank in Austria is innovating on top of open source technologies, to provide transparent data access to consumers. George is the new digital banking arm of Erste Group Bank (EGB). George aggregates all elements of a customer’s financial life, to provide meaningful insight. It can analyse and visualise the way expenditure is split into different elements and align this to a user’s current account. It can also estimate expenditure based on previous history and help users adjust spending if they can see they’re over-spending in some areas.

Launched in Austria in 2015, George now has more than a million users. With George, EGB is taking the step from classic online banking to a digital banking platform; in alignment with PSD2. For many of the new applications and George’s modern user interface, an Elasticsearch cluster is responsible in the background. In addition to the semantic search, this also includes statistics based on Elasticsearch Aggregations.

Erste has also used Elasticsearch to define an API on which the clients (e.g. web front end, mobile clients, specialized clients) are based. This allows the company to present a uniform appearance to customers across all groups and to quickly integrate the developments and innovations of the local banks into the entire group. 

Conclusion

Open source software in the finance industry is now about building on best-in-class open technologies, which have thriving, active communities around them. For institutions with thousands of applications to manage, it is about the ability to transfer and reuse common technologies, developers, procedures, operations and even power-users, across multiple projects. The ability to source talent, and to pull in experience and innovation from other industries that are using the same technologies, adds to this draw.

What is changing is that where this once applied to open development languages, operating systems, toolsets and architectures, this thirst for powerful, reusable technologies is increasingly moving up the stack. Burned by the cost of grandiose one-off big data projects, many companies have now learned to leverage complete open source applications capable of solving a multitude of problems across the business.

To Top