I’m in the afternoon session for the Orange business day as part of the #BlogBus tour. I’m liveblogging this session so please keep renewing the page for updates.
We’re introduced to the relationship between Orange and Xyratex, the latter being a data storage solutions company. Soumya notes that the challenge for Orange was to deliver very high bandwidth streaming content over intercontinental distances to demonstrate their DWDM optical networks. The problem for businesses handling high bandwidth content has always been latency.
Efficiency of network infrastructure is increasingly relevant as social and business networks create ever more data. But when it comes to streaming content, efficiency of video content is crucial in order to ensure the value of content is sustained. Orange conducted a review of available products and chose Xyratex on the basis of their reputation and on their forward thinking lustre file-systems.
Together they built a system designed as a proof of concept for globally delivered CDN content. Key was to ensure that the cost of a 40Gb pipeline would be justified by high bandwidth efficiency.
Eventually Orange was able to easily produce 60% of the 40Gb pipe produced 16 streams of content with half a rack of servers dedicated to rendering the content. All video streams can be simultaneously delivered in HD and with no latency, stuttering or other quality losses.
The data demands of the global business community and the capacity requirements of everything from enterprise infrastructure to consumer devices are growing at a fast rate than existing networks can supply. Just as energy and connectivity issues are now at the coalface of the tech sector, the data storage and transmission, rendering and insights issues are growing at a similar rate.
Orange is not yet considering using this technology as backbone infrastructure but it may develop into one of the solutions for backbone infrastructure eventually – or, as the technology is open source, another telephony carrier may explore that option.
Soumya notes that the true value of the technology is not necessarily increasing the bandwidth and cap ex of new infrastructure but actually maximising the ROI of existing infrastructure. Doing more with small pipelines and overcoming localised caching data storage issues for CDNs. Exobytes of data in a data centre is the ultimate goal.
[JJ’s comment – Australia should really be considering implementing this kind of technology to maximise its global pipeline.]
The next presentation is from Julian Gay on a unified search service for searching across enterprise cloud services.
It is easy for a firm to lose track of content held in a multiplicity of different cloud services. Xen.do is designed as a unified interface to content, as well as to find organisational content, and to act on it.
Product is a mobile-first web experience so that the maximum user base is possible. Initially it provides for CIOs, information about what services are being used by a firm, and it provides an interface for that content. But it is designed, also, to aggregate content for search purposes and to recommend alternative systems for handling content. Employees can sign up to the system and then log in to cloud based systems for real time content. Xen.do does not keep passwords for any cloud system. Only open authentication systems are used to access content held in third party services. It then indexes content for search. So far the cloud services are limited to apps such as yammer, Google docs/drive, dropbox, Office365, Salesforce, etc.
[JJ question: what is Xen.do doing with the content they index? What about commercially confidential data?]
The CIO/admin interface allows organisations to control what services employees see.
At the moment the development of Xen.do is on the Find stage. Next stage is discovery of new organisational data, not currently listed among sources of company information. Finally there will be a recommendation system to encourage companies to become more efficient in employee productivity.
Essentially it is ease of enrolment and free use among employees that will drive organisations to establish business accounts to access insights on employee document sharing.
[JJ’s comment – interesting technology. Should follow up. I have some reservations from a security perspective but I see the value.]
Open Mobile Application Directory – OMAD
The next presentation is on a system to index all the apps held on various app stores. The idea is not just to overcome the problem of finding apps, but also finding equivalent apps on different platforms. The objective is partly to ease the process of moving between smartphone platforms among users.
[JJ’s comment: I can imagine Apple won’t like this!]
Srinivas Chervirala describes the technology as a ‘wikipedia for mobile apps’. Product trials are currently underway for this open platform directory.
Problem is that in application data there is a lot of noise and there is a very poor app discovery experience. The open platform direxctory is designed to assist retailers of smartphones, developers in allowing their apps to be found, and users in ensuring the apps they want are accessible.
The current search platform looks at Android, iPhone and RIM apps. The intention is to also include ovi and Windows apps. It generates data about apps and populates the PeekApp system so that users can find apps that interest them.
[JJ’s comment: obvious value here but might go against business plan for retailers at least.]
Next presentation is on a social local mobile commerce service, MovieNow – a bidding system for movie content. Instead of theatres operating a first-come, first-serve basis for ticket sales, they can regard seats as a perishable inventory, and allow for customers to bid for seats.
Given most theatres only fill 12% of available seats, the bidding system would make the most of high value movie seating but then increase overall sales of seats for lower value (perhaps older or less popular) movie content which they commit to screen but do not generally change seating prices to reflect less popular content.
The business model is based on a business oriented subscription system to the service, with users bidding for free. Theatres can make only a few seats available for bids if they choose, or an entire theatre, for a substantive or limited time period. Users have to guess what discount has been offered. [JJ’s comment: … which seems awkward.]
At the moment, while the app is live, it is only available in Portland, Oregon. The intention is to release to other markets as the concept is tested.
This wraps up the early afternoon session – will be coming back to blog the last session in half an hour.