The Day Bloomberg Plunged Us Into Information Darkness
Bloomberg trading systems are brilliant but – as the April crash showed – they might not pass the ACID test that NuoDB was designed to overcome.
The global financial ecosystem is the ultimate free market, right? With its liquid assets, mobility, lack of barriers to entry and millions of potential traders across the world, you have optimal conditions for capitalism.
OK, not perfect, but as near as dammit, surely?
All you need is for each trader to have the best possible information on each buying and selling decision they make. With all these elements in place, we have the perfect conditions in place so that everyone benefits.
Sadly, that has not been the case.
The information sources for finance traders have been proved less than perfect. That is not a criticism on the quality of the information available to the hundreds of thousands of finance traders around the world. It’s an observation about the delivery. There’s a near monopoly in the market for trading information and, as the recent crash of Bloomberg’s trading service shows, this is a serious flaw.
Aren’t markets meant to be pluralistic? Isn’t their rationale about providing alternatives and best options? Aren’t about markets about flexibility?
When the Bloomberg data and communication network unexpectedly crashed, in April, 2015, it showed how inflexible the market actually was.
That’s not to say that the market was closed, but the light that Bloomberg was able to shed on all aspects of the market, on everything from price information to profile to news, was suddenly shut off. Without that information, the majority of the 325,000 financial professionals around the world, who use Bloomberg’s systems, would be effectively trading blind. Nobody wants to do that, so the failure of the information system created negative sentiment that pushed down stock prices around the world.
Without the guiding light of market news and chat between buyers and sellers, the market was plunged into darkness.
It is human intelligence that makes the markets of Wall Street, Frankfurt, Hong Kong and the City of London hum. When this is such a vital commodity, doesn’t it seem a bit odd to have a monopoly supplier? What sort of market is completely reliant on one source? Bloomberg’s Chat programme is undoubtedly a killer application – by all accounts it’s the hook that inspires so much enthusiasm in its uses. There are other systems around – Reuters offers a comparable level of market intelligence and Money.net aims to compete with Chat – but Bloomberg is the only one with 100 per cent connection.
Perhaps the global finance market isn’t the ultimate market. Yes, money has perfect liquidity and can get around the world twice before a newly appointed CEO has addressed his or her new staff. But information – as the case of the Bloomberg crash exemplifies - hasn’t got the same resilience.
For examples of resilience we have to look to the new generation of cloud computing. The assets of a computing (processing power, storage and memory for example) do have perfect liquidity in the cloud. The global infrastructure of data centers has created a sort of holistic hosting environment, which early adopters have enthusiastically pioneered.
One minute these units of computing production might be rented to a video company user in Amsterdam, the next they might be sold to a gaming outfit in Las Vegas.
In order to consume computing assets like this, the data foundations of global traders need to have matching characteristics.
In database terms, that means creating foundations with the capacity for response. If you have millions of transactions taking place across the world, every second, you need a lot of backup. So – unlike those who rely on Bloomberg – your information comes from several sources.
In database terms, that means having several homes for your data. Replication gives you resilience, and that reserve capacity gives you a robust response to difficulties. Disaster recovery is great – companies like NuoDB are good at that. But disaster prevention is a lot better, and that is a preventative medicine it administers a lot more frequently.
This is where the founding principles of NuoDB’s own version of the perfect information come in. NuoDB has its own equivalents of the free market’s conditions for efficiency. Database vendors like NuoDB also strive to create liquidity, mobility and choice but build on those qualities with an additional layer of assets. NuoDB describes them as scalability, recovery, integrity and consistency.
The technical principles that underpin these foundations, such as administration, sharding and partitioning, need more time to explain.
Suffice it to say, for now, they could give Bloomberg some pointers on distributed deployment and ACID guarantees.