With Amazon recently stating they plan to start offering a content delivery network service as part of their S3 offering and Microsoft reportedly creating their own CDN in partnership with Limelight there seems to be a renewed interest in CDN's. Whether for public usage like Amazons or Private like Microsoft's, the market for pay per use content delivery services is heating up.
According to a recent post on the data center knowledge blog, the Microsoft CDN, which it calls an Edge Content Network (ECN), is designed to beef up the company’s infrastructure to handle “Internet audiences and content offerings that are orders of magnitude larger than today,” according Debra Chrapaty, Corporate Vice President of Global Foundation Services for Microsoft. Chrapaty has said that the ECN will include at least 99 nodes around the world. (For now it looks like that the MS ECN will only be used as part of MS's global infrastructure)
So what else is going on in the CDN world? The answer is simple. Peer 2 Peer (P2P). One of the most disruptive technologies to appear in the world of globalized computing by far is the emergence of P2P technology for NON-consumer usages such as business continuity. Leading this charge are companies like Pando and their unique P4P approach.
I recently met with the Pando team in New York and needless to say I was extremely impressed. Pando Networks describe themselves as a small company on a big mission; to establish a new Internet infrastructure standard for efficient, secure and commercially viable rich media delivery. At the heart of Pando's platform is its "P4P" technology.
For those are unfamiliar with P4P, or Proactive network Provider Participation is a method for software to optimize global peer-to-peer connections. P4P proponents say that it can save an it users significant costs, and that using local connections also speeds up download times for P2P downloaders by 45%.
Pando describes their system as a Hybrid P2P platform that leverages both P2P and HTTP protocols (Peers and Servers) to optimize content delivery for performance and cost efficiency. The general ideas is as your content becomes more popular, bandwidth is transparently offloaded from your CDN Servers to your consumers ("Peer Cloud"). As your content moves up and down its consumer demand curve Pando intelligently shifts between sources and protocols in real-time. What's great about P4P is it would seem to be a logic fit in existing cloud storage deployments which could include Mosso's CloudFS or Amazon S3. (Seed your CDN from the cloud, let your users do the rest.)
The big question is why hasn't Akamai done more in the emerging cloud CDN space? I think the simple answer is because they don't have to. Akamai seems to think that it's easier to let others battle it out for their left overs. I would guess they are waiting for the right time when they will most likely just a acquire a few strategic startups, entering the space when they're good and ready. From my point of view, this is a rather dangerous way to handle technical innovation and R&D. Rather then lead, they seem happy to be a kind of "fast follower". The problem with this approach is they may end up completely left out, a relic.
As we continue our long march from the desktop to the cloud, management of digital assets on a global basis will be a key component. CDN's are well on their way to becoming one of the more important aspects of this big shift toward the cloud.