January 2, 2012

2012 - The Year of the Cloud

It may come as no surprise to anyone that I do in fact, heart the cloud.  Having spent the passed few weeks reading the most excellent book ‘The Big Switch’ and generally reveling in my post Proof of Concept week at Microsoft (to be epically blogged about later), I am convinced that 2012 will in fact be, the year of the cloud.  Here’s what I mean:

1. The computer as a Utility

This one is pretty easy to recognize if you have read The Big Switch – the economies of scale are about to arrive at where they need to be to obviate the need to even have a desktop computer.  They are already there with regard to online storage.  What a disruptive market!  I see 2012 as being the year when 100 Gb of storage for $100/year becomes not just a reality, but the norm.  Where companies will differentiate themselves remains to be seen, but the money is where you find full feature sets with what you can do with those files/data – be that music (a la Amazon or Google with their 20,000 free storage) or raw files (a la Google Docs or Azure Blob Storage, DropBox or Box, etc) or the raw computing power that is starting to take off in the cloud (a la Azure).


2. So many tools, so few things to stand apart with

The biggest challenge of the day is currently what to do with all of the ‘free’ stuff.  I want the space of Google Docs (I spent $5 for 20 Gb), the toolset of DropBox (multi-computer plus tablet synchronization), the functionality of SkyDrive (Office Web Apps are quite simply – amazing) and the simplicity of iCloud (it literally freaks me out that all my stuff lands everywhere I am without me actually doing anything).  The problem is the fragmentation, so I have come up with a few things to keep things where I need them – see next entry.  Again, the differentiation is key.  Where companies provide the raw form of compute and storage – Azure and to a lesser extend Amazon – these will be the foundation of all things to come.  The companies (and hopefully open source gang) should be able to leverage these foundational components to build truly fantastic resource providers.

3. Need compute – will travel

I ran into a friend the other day, and he told me the most amazing story – one that I have told myself before, but is always great to hear additional validation (numbers are relative).  The story goes something like this: Company A has 2 million dollars worth of idle computing hardware (in this instance, running a giant HPC cluster) to perform a calculation once per year.  The calculation takes their 2 million dollar machine 3 months to complete and is completely non-interactive (batched stuff).  The other 9 months out of the year, it sits, literally idle.  Company A engages Microsoft to rock a POC whereby they ramp up 400 Medium size HPC compute Azure instances, process takes 6 HOURS to complete and the result is then delivered.  After that, the compute instances are torn down.  Total cost? ~$1,500.  2 Million in capital investment vs $1.5 Thousand in operational expense.  3 Months to process the result vs 6 Hours to process the result.  Case. Closed.

Now – how do we apply that to the end user?  With the ubiquity of broadband (not so much in the good ol USA cause we apparently SUCK at that topic, but perhaps South Korea), such computational excellence can be had for literally pennies a day for the average consumer – and the costs continue to plummet.  Add to that reality the availability of excellent traveling hardware (currently the iPad 2 but soon (please for the love of all things holy) the Windows 8 tablet).  This form factor carries enough onboard to perform the basics – when you need to go with higher processing umph, enter the cloud – log in remotely from that tablet or any other computing device with your one name and password and there before you sits Windows, your files, your apps, everything, everywhere – this will happen this year — I hope.

It’s going to be a great year – I cannot wait to get my hands on the truly amazing products coming out into the marketplace, and hopefully, getting closer to not needing a standalone machine to lug around.