Google Cloud Spanner – the next big transformation in financial services?


Cloud database technologies continue to evolve at pace – handling huge amounts of data, scaling horizontally and not costing the earth.

With the launch of Google Cloud Spanner, banks can now harness horizontally and globally scalable data stores across their entire trading business, creating dramatic improvements in cost efficiency, resilience and the ability to manage complex global trading needs. Whilst massive horizontal scalability had been the preserve of NoSQL data stores, Google Cloud Spanner brings SQL and transactional relational data into this arena.

At an upcoming joint event, Google Cloud Spanner engineers will be unveiling the full capabilities of the new service, whilst GFT specialists will be demonstrating real-life use cases using Google Spanner, to significantly transform how globally developed trading applications can be deployed for Client MIS, Risk, P&L and reporting across asset classes such as Equities, FX, FI and Commodities. To join us for the event, please click here.

Google Cloud Spanner offers a cost-effective approach to managing worldwide scalable data for the world’s largest financial institutions and this event will deliver great insights into the exciting opportunities this revolutionary technology can create.

Current challenges of existing technology

As trading systems have become faster, more automated and widely distributed, this has brought increasing complexity and headaches to the IT teams of many financial institutions as they battle to ensure localised datacentres can scale to meet the challenge of increasing data, are resilient in a crisis situation and are inter-connected to enable firms to have a global view of all their trading activities.

Typically for trading firms, each region will have a stack of high-performance on-premise servers and a complex array of software products to support large numbers of transactions – particularly for peak trading events. Outside of peak trading events, this stack of technology investment is significantly under-utilised. Order Management Systems (OMS’s) are built in-house, or purchased to handle this trading – in asset classes such as FX, Equities, Futures, ETFs, Fixed Income or Commodities.

Most firms currently use database technologies such as Oracle or SQL Server for their OMS database with, for example, a cluster of servers in London, Tokyo, New York etc.; in effect creating a global trade data store. The requirement is to store every step through which the order passed, for business reporting and regulatory reporting purposes. In a globally integrated setup, if one server crashes it can often be a huge headache to ‘resync’ that server whilst trading continues and more data flows into the system. The end result is a highly complex set-up of local clusters on one site, often with localised back-ups and caching, and attempts to duplicate data across multiple global sites; creating a huge challenge to keep the data consistent across all servers. Those firms that have attempted to have one global view of all their trading data, even in just one asset class, will attest to how expensive and complex the solution becomes. In fact, the expense and complexity are so high, that many firms do not attempt to consolidate trading in real-time, but may in fact use batch or T+1 processes.

One of the key challenges is that of attempting to provide position and risk reporting across the large Oracle or SQL Server databases and delivering this on a global basis across all centres. Often with trading taking place in multiple time zones it can be many hours after the closing of a trading region until the data can be aggregated with other trading venues to provide a global view. Banks are spending a fortune on trying to get a global Risk or P&L view and are struggling with the enormous data stores – expending more effort on keeping the system up and running, rather than building the analytics and insights that modern trading operations now require.

How can Google Cloud Spanner redefine the challenge?

The killer feature of Google Cloud Spanner is its ability to run SQL on relational data that is globally and ‘transactionally’ consistent, and its ability to scale (or span) horizontally across data centres and regions. Other horizontally scalable databases do exist, but they do not easily run SQL, the mainstay of trading and reporting systems that deal with enormous flows of data. Whilst any migration to new technology will require platform re-engineering, with Google Cloud Spanner there is the advantage of the familiarity of SQL and the relational data model, whilst enjoying the interconnected and scalable nature of a system that can scale up and scale down on a global basis as the need for capacity changes. For example as trading venues open and close, additional servers can automatically come online or be switched off as and when required, something that is impossible to do with fixed and established Oracle and SQL Server systems. This is the promise of the ‘Elastic Cloud’ in practice.

Given the global and connected nature of these systems, there is no longer the need to build by hand the complex synchronisation and replication of data – all of this is built into the Google Cloud Spanner product, allowing connectivity and resilience on an unprecedented scale. In addition anything that needs a global view, such as P&L, Positions, Client MIS, Risk, etc., is much more easily built. Imagine being able to have a global position reported within minutes of each trading region having closed? The benefit for the business could be enormous, whilst at the same time the cost savings of reducing the required capacity in any single location as it goes offline would dramatically improve the bottom line.

The speed of making change is another key benefit. The simplicity of pressing a few buttons to spin up a new global database capacity is unprecedented. Even changing database schemas whilst daily trading continues without a pause is possible. Historically such changes could take many months (or more) in previous relational-based systems.

Benefits of Google Cloud Spanner include:

A significant reduction in complexity
A significant reduction in cost
A significant reduction in management overhead and improved agility
A significant improvement in availability and reliability
The ability to deliver advanced analytics over huge data sets

Other use cases using Google Cloud Spanner

Whilst we have focused our proof of concept on trading, there are many other use cases for such a connected global and expandable system.

Many people who begin to experience Google Cloud Spanner comment on the ease of use, the connectivity and scalability of a truly global connected database system. Once they experience these benefits, users often come up with new use cases and opportunities that are now possible for the first time; literally being able to create a cost saving or a revenue generating activity that has just not been possible before.

Users soon begin to realise that having a low latency global view, and the ability to rapidly scale up and scale down according to the database requirements, Google Cloud Spanner could well provide the biggest transformation yet in financial services!

There is still time to register for the upcoming Google Cloud Spanner – A revolution for Financial Services event in London by clicking here.