jump to navigation

The New Asset Class December 14, 2009

Posted by James Webster in : finance, virtualization, development , add a comment

I touched on this idea briefly before in Commodity Markets for the 21st Century and it looks like the idea is evolving further…

In Hedging Your Options for the Cloud Joe Weinman discusses how as cloud computing turns computing power into utility, its pricing and risk management strategies around volatility of prices will begin to reflect those of other commodity markets (electricity in particular).

As if in response to the previous article, Werner Vogels of Amazon announces ‘Spot Pricing’ for EC2; customers requiring EC2 compute time at some point in the future without hard deadlines can nominate a maximum price they are prepared to pay. When the spot price (defined in financial markets, particularly commodities and FX, as the price for immediate delivery of an asset or commodity) drops below this level the customer’s EC2 instances will spin up and start computing. When the spot price climbs above the customer’s specified level the EC2 instances will automatically shutdown. The latter will require the workloads to be resilient to abrupt termination but this should be fine for robust batch jobs.

So I wonder if futures/forwards/options markets may grow around this new commodity one day. Probably not at the moment as there is no mechanism for selling short. Also unlike deregulated and competitive electricity markets there is only one ‘generator’, Amazon. However the EC2 APIs are published and in fact the Eucalyptus project is building an open-source implementation of the EC2 APIs. Eucalyptus has already been integrated into Ubuntu and RightScale’s toolsest for virtualization. So it might seem that the barriers to entry for a potential Amazon competitor are quite low, however Amazon have already achieved significant economies of scale (which will increase further with their planned entry into the Asian market).

It would seem natural for Microsoft or Google to challenge Amazon directly by providing their own implementation of Amazon’s EC2 APIs, however their own cloud computing offerings require developers to code to more proprietary frameworks (although there are open-source efforts to implement the Google App Engine on EC2; AppScale being seemingly the most advanced) rather than EC2’s approach of just providing access to a virtual machine. Maybe they have already run the math and decided that competing directly with a low-cost provider like Amazon is like wrestling with a pig… you both get dirty and the pig likes it :-)

It would be great for competition and innovation if cloud computing was more transparently portable between providers, and the resulting fungibility would benefit consumers of cloud resources in being able to manage their operational expenditures.

Xignite November 12, 2009

Posted by James Webster in : finance, development , 3 comments

Xignite, ‘on demand financial market data’ recently got some coverage from TechCrunch regarding their upcoming integration with StockTwits and iPhone portfolio management apps from Turing Studios.

Xignite have some great looking APIs and I understand that their business model is all about providing the data ‘on demand’, but it would be great if they could combine their use of open APIs with a basic market data streaming offering as well. They could take the ‘comet‘ approach, use a 3rd-party technology like Caplin Liberator or Lightstreamer, or perhaps the best forward thinking option, adopt HTML 5 Websockets. Since they would be delivering their stream via the Internet rather than leased line or private packet switched network there would be more latency compared to a professional solution such as Bloomberg or Reuters, so is there a market for streaming financial data at a lower SLA?

I would like to do some experimentation with these APIs, maybe amCharts’ WPF Stock Chart control will be an excellent way to visualise the data. Hopefully amCharts are working on a Silverlight version as well.

Tis the season for consolidation August 25, 2009

Posted by James Webster in : finance, development , add a comment

Tibco buys Datasynapse, which is pretty big news for the City since most derivatives houses probably have one or the other. Will this sort of consolidation at the big end of the market scare some architects off and make them focus on open-source players such as GridGain? Are there any grid deployments running GridGain for production intraday & overnight risk at a bank?

Open-source is no stranger to the consolidation trend either;

Terracotta acquires Ehcache: from a capital markets perspective, Oracle’s Coherence (which itself arised from acquiring Tangosol) seems to be the 500lb gorilla in the ‘in memory data grid’ space. The combination of Terracotta and a deeply integrated Ehcache might be an appropriate alternative solution for the primary use case of Coherence in bank deployments; a market data/trade/position/risk cache & time series database.

VMware acquires SpringSource which itself acquired Hyperic: VMware is gearing itself up to be a major cloud player by offering a complete virtualised Java development stack, after all isn’t the recently announced CloudFoundry Java’s answer to Microsoft and .Net’s Windows Azure? Is a database acquisition or purchase of Splunk that far away?

Co-operation rather than consolidation; the open source trading platform Marketcetera has established close relationships with Sky Road and NYSE Technologies’ to provide a hosted solution for Marketcetera. Given Marketcetera uses Spring extensively some involvement with CloudFoundry is a possibility as well.