Telcos in the Cloud: IBM, Orange, and Oracle @ Telco 2.0
This afternoon has been an optimistic one at the Telco 2.0 executive brainstorm - you could say our heads were in the clouds. Specifically, the Cloud Services session has been more than interesting. Speakers from IBM - Craig Wilson, VP Global Telecoms Industry - Orange - Stephan Hadinger, Chief Architect for Cloud Computing - and Oracle, with Neil Sholay, VP Cloud and Communications - gave a string of presentations that encouraged us to hope that the cloud might not be quite as airy and insubstantial as it might sound.
Wilson thinks operators have a "unique market opportunity": they have key assets like end-to-end control of network resources, an existing enterprise customer base, name recognition, and channel marketing expertise. We'd add to this that the ex-incumbents tend to have a footprint of network and property assets that you need if you want to distribute cloud resources, for reliability, and also bring them close to the end user, for low latency and high performance.
Similarly, Stephan Hadinger pointed out that most cloud providers can only provide network infrastructure services (load-balancing, anycast DNS, spam filtering, DDOS protection, content- and applications-delivery networks) as software, when really high performance requires a presence physically located in the right place.
Wilson went on to point out that "dev/test clouds" were a key use case - that is to say, it's just easier to do rapid-prototyping of a new idea if you can check out an Amazon EC2 machine image, or use Google App Engine, than if you have to negotiate with the IT department first.
Interestingly, Neil Sholay of Oracle said something similar. He pointed out that Oracle had started to "get" the cloud after it was discovered that their developers were in the habit of using public clouds rather than going through channels. At least, he said, this was better than some of their customers, who had discovered their critical cloud applications when they were shut down after the inventor left the company and stopped paying for them.
As a result, they had now invested in internal cloud capacity - often, applications would be prototyped in the cloud and then pushed to the traditional infrastructure.
More broadly, he pointed to a case study of a major Oracle customer - a bank - that had achieved cost savings of more than 30% in both operations and innovations. Very interestingly, only 19% of the savings came from the core "cloud" project of data-centre virtualisation - much else came from better practices in innovation, operations, maintenance, and the like.
The amusing thing about these presentations was that they track very closely with the content of an as-yet unreleased Telco 2.0 analyst note on the potential for operators in the cloud. We thought rapid prototyping and scaling of new services, IT cost control, and geographically optimised CDN/ADN were the key elements. Watch this space!