Independent research by Grey Spark supports OpenMAMA adoption

Nov 29, 2016

Replumbing the Market Data System

4 Nov, 2016. By Jonathan Chambers — Analyst Consultant, GreySpark Partners. View the original article here.

This article examines the OpenMAMA open source initiative and its place in replumbing market data systems, helping end-users to innovate, reduce vendor lock-in and conquer rising costs.

The complex network of separate pipes that must be installed and maintained by end-users to receive real-time market data is helping drive the global spending on this information skywards. This article examines the open source initiative, OpenMAMA, looking to disrupt the market data marketplace by replumbing the market data distribution system and help end-users conquer the rising costs of this information.

An intensifying thirst for a greater volume of real-time market data by financial institutions is driving prices for the information upward. However, the current plumbing of the financial marketplace’s market data system weakens any cost reduction strategies as end-user applications are dependent on incumbent market data sources. Vendor lock-in reduces the bargaining power of market data end-users in financial institutions and leaves the vendors holding the upper hand in contract re-negotiations; a position from which they can freely raise market data prices. To overcome this disproportionate supplier power, GreySpark believes that financial institutions must replumb their market data systems so that their architecture is vendor agnostic.

An Inefficient Market for Market Data Services

Market data costs remain one of the largest outgoings for financial institutions, with global spending rising from USD 22.7bn in 2009 to USD 25.6bn in 2015. A maze of proprietary piping currently funnels market data from vendors to end-users, with vendor-specific technology components required to consume multiple vendor feeds. Specific application programming interfaces (API) must be used to allow for the delivery of vendor’s data to each application that consumes the data to take into account the vendors’ proprietary semantics and data models.

Vendor lock-in arises, primarily as a result of the cost, time and risk of changing from one vendor’s API to another for each end-user application that connects to the new vendor. This often leaves applications tied to incumbent vendors’ technical and data infrastructures.

Institution-wide market data architecture currently resembles a complex system of siloed market data mechanisms designed to meet specific end-user data requirements that differ by asset class, geographic coverage, latency and type of use. Installing and maintaining multiple pipes carries high IT support and maintenance overheads and in the large majority of cases delivers no direct competitive advantage to the financial institution. Furthermore, costly re-piping must also be undertaken when applications and distribution platforms need changing or upgrading.

Vendor lock-in fetters market forces, enabling incumbent market data vendors to artificially ratchet up prices and prevent would-be challengers from competing on a level playing field. As a result, end-user applications are often left consuming market data at exorbitant prices. In addition, the non-negotiable bundling of incumbent vendor’s market data packages leaves financial institutions paying for data that they may not use. For example, Figure 1 demonstrates the rising costs to industry participants of vendor market power, showing that between 2005 and 2013, Bloomberg increased the cost of its terminals by over 2% per annum on average. Vendor lock-in drives inelastic pricing at an end-user level, despite the raw market data being a highly substitutable good.

Figure 1: The Rising Cost of a Single Bloomberg Terminal (USD)

Source: GreySpark Analysis; Quartz

Disrupting the Market Data Marketplace

Aiming to disrupt the status quo and replumb this inefficient system is the Open Middleware Agnostic Messaging API (OpenMAMA) initiative. This open source initiative is hosted by the Linux Foundation and commands powerful industry backing from JPMorgan and a number of other Tier I investment banks. OpenMAMA is an API standard that provides an open source code to create an abstraction layer that removes the dependence of end-user applications on vendor-specific APIs, thus enabling market data-consuming applications to take data from with multiple vendors’ distribution platforms.
OpenMAMA provides the source code for a low-level API capable of transferring the simpler structures of level 1 market data. The market data-specific overlay, Open Middleware Agnostic Market Data API (OpenMAMDA) provides a consistent abstraction for the objects of level 2 market data, such as order books.

Despite these laudable initiatives, financial institutions must still decipher the various different symbologies and data models used by market data vendors. To fully address the issue of proprietary technologies, a non-software standard open sourced market data model for mapping raw market data would also be required.

Utilising an open market data model alongside the market data-specific API abstraction layers of OpenMAMA and OpenMAMDA provides end-users with true vendor agnosticism. This target state leaves end-user applications unreliant on proprietary symbologies, data models and APIs. As such, a vendor agnostic architecture transfers bargaining power away from market data vendors toward end-users and enables Tier II market data and technology vendors to freely compete with incumbent vendors. Increased competition within this market will enable end-users to tackle rising market data costs, either through leveraging cheaper Tier-II vendors or driving harder bargains in contract re-negotiations.

Less stretched IT budgets can also be re-focused on delivering competitive advantage, rather than being soaked up by the development and maintenance of non-differentiating software.

Open Source Development

Recent trends among a variety of different types of market participants to pursue utility or syndicated models for technology utilisation represents a marked increase in shared industry development; for example this trend is becoming prevalent in non-competitive areas such as regulatory compliance. Additionally, as open source initiatives, these standards will enable industry-wide collaboration to reduce costs through the sharing of bug fixes and added functionality.

A 2016 Rackspace survey found that 63% of financial services’ respondents stated cost savings as a driver of open source software adoption. Open source software is downloadable with permissions to redistribute and modify it freely and is thus more flexible than proprietary software. Furthermore, the initiative’s open source license, GNU LGPL v2.1, is appropriate for commercial applications, as modifications to the code can either be retained by the financial institution or contributed back to the initiative.

Looking Into the Future

GreySpark Partners believes that institutions will increasingly adopt the modular, vendor-agnostic architecture provided by the OpenMAMA abstraction layer, and the increasing number of market data cost reduction initiatives underlines the growing industry demand to challenge the status quo. For example, the increasingly popular messaging initiative Symphony is backed by large asset managers and Tier I investment banks as a low-cost, cloud-based messaging application to tackle reliance on Bloomberg chat. Also, Money.Net is a software-based market data application aimed at challenging the need for owning the hardware of a Bloomberg terminal.

Insulating end-user applications from the source of the market data will enable end-users to hold a data-centric, not vendor-centric world view, and align market data delivery more closely with evolving business needs. This can allow businesses to take advantage of innovative services provided by Tier II market data vendors that would have previously been uneconomical to consider.

While financial institutions may be deterred from replumbing their market data systems by the large up-front cost and bank of resources required to achieve it, vendor agnosticism has the potential to deliver substantial long-term cost savings. Furthermore, as the critical mass of users increases, the quality of the APIs and available connections will grow, as will the commercial support available for the open source project.