About twelve years ago – by lack of available CDNs and CDN technology – we had to develop our own basic CDN technologies, to be able to produce our large-scale webcasts for pop festivals, events, enterprises and governments.
Our requirements back then (1996/1997) were quite simple:
distribute assets and live stream relays between geographically setup streaming servers
Recognize users by ISP via their IP address and relay them to a matching server
Gather and merge log files for centralized analysis and view reporting
In 1998/1999 we added two additional requirements:
generic support for multiple streaming services (back then RealServer, Windows Media Server and QuickTime Streaming Server).
maximum density to achieve maximum flexibility and low cost deployment (we developed a technology called parallelization so we could run multiple services on one box).
In 2002 I started Jet Stream BV because I foresaw the need for commercial CDN technologies. We rewrote our technologies with these requirements:
Make the technologies more generic to make them commercially attractive
Write documentation so third parties could design, deploy and manage the platforms based on our technologies
Add more dynamic features and make the technologies more robust
Add user management so streaming platforms and CDNs could be used in a multi-user environment
Over the years we thought of many new features and got great feedback from customers. Between 2003 and 2007 we managed to license the technologies to a wide variety of customers, mainly in the BeNeLux region, and we added many great and unique CDN related technologies and features to our VideoExchange (CDN) and XL Media Server (parallelization) products. We also launched our own StreamZilla CDN, based upon our own technologies.
We were quite successful. We managed to grow our business without investors, loans or debts. We had been profitable every year and grew almost 100% every year.
In 2006/2007 we decided to move to a global approach. We wanted to take our technology to a next, cutting-edge level. So we wrote down some requirements:
A modern CDN should focus on video delivery since that is where all traffic growth – and the pain – and the business – is.
A modern CDN should be cheap to design, deploy and operate.
A modern CDN should SAVE telcos money instead of cost them extra.
A modern CDN should allow telcos to offer premium services to their customers.
A modern CDN should support various licensing models (licensing, leasing, pay per TB, pay per subscriber, pay per user)
A modern CDN should support various white labelled, wholesale and reseller business models.
A modern CDN should stick to open standards where available / where possible on all levels.
A modern CDN should introduce new open standards where other standards lack or are missing.
A modern CDN should have open standards based APIs to integrate with 3rd party delivery technologies, appliances and servers.
A modern CDN should offer open standards based APIs for customer workflow integration.
A modern CDN should offer open standards based APIs for inter-CDN integration.
A modern CDN should be designed and documented so it can be deployed by third parties without extensive training or explicit skills (OEM, system integrators)
A modern CDN should be designed and documented and automated so it can be deployed in a matter of weeks by average skilled engineers.
A modern CDN should be so easy to operate that it can be operated by a very small team of average skilled engineers.
A modern CDN should generically work with all market-accepted 3rd party delivery technologies for internet, mobile and IPTV streaming & asset delivery in a hybrid way.
A modern CDN should guarantee 100% compliancy with the market-accepted delivery technologies: rewriting or mimicking would be too expensive and would be a liability in terms of future-proofness.
A modern CDN should be able to maximize throughput and density per delivery server, per PoP and overall.
A modern CDN should be able to run on commonly available and affordable hardware and various operating systems and should not rely on expensive black-box third party appliances.
A modern CDN should be flexible to support the storage versus bandwidth / resources costs calculations.
A modern CDN should not rely on third party systems or technologies for asset replication (no P2P!, no caching!, no syncing!)
A modern CDN should not need to rely on third party systems or technologies for user redirection (no DNS!, no BGP!, no global load balancers!)
A modern CDN should be able to use both common off the shelve hardware that we can tune, and appliances, in a hybrid way, to support various cost / performance benefits
A modern CDN should not need to rely on virtualization technologies for deployment and density.
A modern CDN should support various storage solutions (DAS, NAS, SAN, etc) in a hybrid way to offer the most cost effective and scalable solution.
A modern CDN should be built as a framework, with logical layers and logical building blocks, for maximum flexibility, scalability, future enhancements, and 3rd party integration.
A modern CDN should not be locked into the network layer but should support deployment in any IP network, and should allow to span a CDN over multiple IP networks (such as the Internet) as well.
A modern CDN should be able to fit in all common architecture designs, with core, redundancy, bursting and edge building blocks.
A modern CDN should be able to run for various businesses like telcos, hosting providers, streaming providers, enterprises, broadcasters.
A modern CDN should not impact the underlying infrastructure: no network or devices replacements, migrations or upgrades.
A modern CDN should have (near) realtime, independent, popularity monitoring for any asset and stream.
A modern CDN should have (near) realtime, independent, dynamic asset replication / relaying for any asset and stream type.
A modern CDN should have (near) realtime, independent, dynamic and transparent geo user load balancing.
A modern CDN should have (near) realtime statistics per asset, per user and overall.
A modern CDN should be able to accept realtime overload, QoS and incident triggers from the underlying delivery and network resources.
A modern CDN should not have intelligence at the edges but at the central facility because edge intelligence means extra complexity and extra costs.
A modern CDN should offer dynamic scaling, clustering, load balancing and migration options.
A modern CDN should allow to manage dedicated, shared and bursting resources per customer, plus allow to manage bursting thresholds per customer.
A modern CDN should offer a single web dashboard for the administrator(s) so they can oversee and manage the entire service in an integrated way.
A modern CDN should support management of commercial packages and products.
A modern CDN should support a full-featured and automated customer provision system.
A modern CDN should offer a detailed NOC with uptime, load, status, replication and redirection reports.
A modern CDN should support export of traffic and product price details per customer for billing purposes.
A modern CDN should offer customers a full-featured self-service dashboard to save on customer support and to offer a professional service.
A modern CDN should offer customers easy asset uploading, management, publishing, live stream setup features.
A modern CDN should offer customers full-featured realtime statistics per individual asset and overall for the account.
A modern CDN should offer customers full-featured historical reports for detailed consumption analysis.
A modern CDN should offer customers a full-featured set of open-standards based API’s for content ingest, workflow integration and account control.
A modern CDN should be secure and safe to use.
A modern CDN should have auto recovery and manual restore features.
A modern CDN should first of all be a horizontal distribution/delivery service that allows any CDN user to integrate their own vertical workflow.
A modern CDN should be able to scale in terms of throughput, #hits per second, #new assets per second, #new GB’s per second, #servers.
A modern CDN should be written in a modern software language for faster development and deployment of new features.
A modern CDN should be able to be integrated with other CDNs in a CDN overflow way.
A modern CDN should be able to be integrated with other CDNs in a CDN overlay way.
Every individual requirement was discussed over and over. We researched and tested various scenarios, technologies and options, to be sure that we could come up with the most future-proof, stable, scalable and manageable solution. We have invested a lot of time and money in what we call this True Next-Gen CDN technology. We reused a lot of existing technology, but a lot of code was rewritten. Virtually all requirements have been implemented by now. And there are many new features and improvements on the road map.
I am extremely proud of this and I am extremely proud of my dedicated team: we have put our heart and soul into this: it took us a lot of headaches, blood, sweat and tears. VideoExchange has become a greater product than we imagined when we started. We managed to squeeze so much extra smart stuff and features in there. It is all 100% unique IP that covers 15 years experience with streaming and 12 years CDN technology development, which is more than most CDNs can build on.
Late 2008 we showed a sneak preview of ‘VideoExchange CDN Suite 2008′ at Streaming Media Europe, London. In March 2009 we officially announced the software. The feedback was awesome. People from other CDNs could just not believe this. Industry analysts were very positive. The features are industry leading. The management interface blows away any competition. But it wasn’t the technology they liked. It was our clear vision and market approach. We still think as a content owner and understand their needs. We analyze the industry, the value chain and we understand the telcos issues and strategies. That knowledge is translated into technology. Technology does not dictate but it follows a vision. I think that this sets us apart from other CDN vendors. We don’t try to sell a box, but we sell a vision that (hopefully) matches to the telcos strategies. We are not forced by investors to focus on short-term features and marketing blah. We do this because of passion, expertise and enthusiasm. The technology is just a tool that bundles this all, it is an enabler.
Today, eight months later, we have already licensed our next-gen CDN technology to seven customers who have deployed various CDNs from Russia, Latin America, Europe, to the USA. More to follow. We have launched a rebranded StreamZilla CDN, built upon the same technology. And before the end of this year we will have setup a pilot CDN overflow/overlay project in cooperation with some great customers. So we did not only manage to develop a revolutionary (some call it industry disruptive) technology, we have also been able to market and monetize our efforts. I haven’t seen any other CDN solution this mature and feature-rich so far. I haven’t heard of any CDN vendor who has managed to get a new product out there so quickly, let alone be profitable from CDN technology. So it looks like we are on the right track, and yes I am very enthusiastic ;-)
Let’s meet at CDN Strategies Summit or Streaming Media Europe and discuss your CDN opportunities! Oh, by the way: if you have specific CDN related requirements: let’s discuss them and put them on our road map. Do you have another vision? Share!
No comments yet.