Content on the Edge

The Edge again and Content Delivery.

The Edge concept was something I noticed a few weeks ago when I got sucked into YouTube one night watching videos on major data centers.   YouTube's suggestion engine is pretty damn good.  And scary.  I went from Mighty Car Mods to DJ Tech Tools to Jay Leno's Garage to Windows Azure Data Centers, the 'Long Tour', along with Inside a Google data center.  Interesting side note, check out Microsoft's Project Natick.  They are building underwater data center pods..  Crazy.

Anyway..  What I got thinking about Edge and content delivery.  Some split this out into Content Delivery Architecture, but it falls inline with Data and Information Architecture.   As always I seem to lean towards the more technical side of things.  So when looking at the Data Architecture and how things are structured and kept and delivered one has to be aware of the last part.  If a large part of your business strategy is with streaming media or some type of content that is delivered, like downloads, etc, then how that delivery can make a big impact on the over all Enterprise Architecture.
I once knew someone that worked for a large animal activist organization and they did a bunch of streaming content of which a lot went viral or at least had some significant world wide exposure.  Knowing what I did about their infrastructure and their data center I was always curious about how they were able to handle the massive amounts of bandwidth required to deliver that content.  The answer was surprisingly simple, Content Delivery services.  Akamai Technologies was one of the originals and is still a large player but Google and Microsoft are also key players in the arena now.  The edge part comes in with where this content is pushed out from.  Pushing the content to major "cross roads" of the internet can have an impact on costs from a few different angles.  Lower bandwidth costs and lower equipment costs are the easy ones to think of.  But there are some of those hidden costs, like DDoS protection, that don't immediately spring to mind.  There is also the costs that would be associated with manpower, maintenance, support, etc, etc.  All these things are important parts of Data and Information Architecture and Enterprise Architecture as a whole.

An interesting example of this was from ProtonMail the other week.  ProtonMail is a free, end-to-end encrypted email service in Switzerland and they got hit with a MASSIVE DDoS.  ProtonMail takes encrypted email seriously, very seriously.  Knowing their data model and business proved to be a challenge for them in selecting a viable DDoS defense system.  They have a really interesting Blog post, Guide to DDoS protection, that outlines the measures they took to not only protect themselves from further DDoS attacks but to also protect their business model of providing true and complete secure, end-to-end encrypted email.

I'm still researching more on the practice of the edge really because it's something that I don't deal with in my normal work.