From The Expert Feature Article
October 27, 2014

Cloud: Nothing In A Box Ever Rocks!


The $64B question: Is cloud in a box a game changer for Microsoft (News - Alert)?

Answer: Today? Yes. Ultimately? No! Why? In the end, no one will have hardware or software onsite because we are an increasingly mobile society, thus onsite cannot compete with off-premise clouds’ massive bandwidth, insanely low cost models, certified premium support, and its ability to quickly be available, as well as fail between multi-geographic regions.

Unless you’re a dreaded cube dweller with no hope in sight of ever escaping the maze, then having resources in an office (whose infrastructure includes everything from power to security), bandwidth is no match for a “real” cloud – then it’s really no advantage at all. 

Advances in AnyCast networking (which is very popular amongst web & DNS providers), along with Distributed Object Storage (RedHat bought InkTank CEPH for $175M in May 2014 for a reason) spell DOOM in big bold letters to this Microsoft, VMware, (your big vendor here), hybrid cloud fantasy.  

Advice: If It Pops Out Of A Box (News - Alert)? Whack-a-mole It!

I remember when Citrix bought Kaviza and offered “VDI in a box” back in May 2011.  It didn’t go anywhere. Yet, Microsoft will post huge initial numbers on what is basically “Servers in a box” for a few reasons:

  • It’s a lot easier to do
  • Customers are clinging to onsite infrastructure like a teddy bear – it fills an emotional, but not a logical or financial need
  • Microsoft on-boards 10,000 customers a week to its cloud so it has huge marketing and momentum

That said, long term, despite early big numbers and wins, this is a huge failure. First off, why introduce hardware into the mix when Microsoft is a software company? Their strategy is to build bit-level synchronization into every windows operating system so you have a “clone in the cloud” – the hardware is just smoke and mirrors to get to you to do this.  Imagine your drive fails at 2am and you have to let the Microsoft guy in. Seriously? No thanks man, this is why we moved to the cloud to begin with!  Letting the telecom guy in was dangerous enough back in the day - letting Mr. Blue Screen loose in your corporate dojo is pure insanity!

Future Cloud:  The World in My Eyes (Cloud CTO 101)

The speed of light over glass problem is only solved with distributed object storage technologies like what dinCloud uses with InkTank CEPH. Think about it: only when your servers, desktop, storage, and everything else lives as bits, which are replicated everywhere worldwide, is it ultimately highly available at a quality of service which matches local access. The goal is this: you don’t want a bunch of power sucking junk in your office but you do want a customer experience that feels like the resources you’re using are so close they are whispering sweet nothings in your ear all day.  So what’s the best day to do this?

I see data living in a “2 local, 8 remote geography” model in the future thanks to distributed object storage systems (all clouds have these now, for a reason). Meaning, you’d put your server, desktop, and storage in the cloud in Los Angeles, and it’s replicated to another nearby city (e.g. Irvine, San Diego, etc.),  but then 8 additional copies are kept in Chicago, New York, London, Dubai, Beijing, Sydney, Rio de Janeiro, and Mexico City.  This is how AnyCast DNS systems work today. They are both impervious to denial of service attacks, but also offer geographic quality of service by always responding from the datacenter closest to you.

If you combine AnyCast networking along with Distributed Object Storage you have CLOUD NIRVANA.  This combination gives you blazing fast local access to your resources with a follow-the-sun access and failover strategy which cannot be beat.

Technologies already exist and are beginning to mature where you can run the same server active/active in multiple geographic regions off a unified set of distributed object storage bits such that possibilities are limitless. You use a virtual desktop in Los Angeles, but fly to London? Bam, same performance, desktop, and data bits – it’s yours but local to wherever you go.  Users in Los Angeles and London on the same server cloned yet synchronized at both their sites in real time? Bam, done. You think a box can compete with THAT? 

Takeaways

My bet is that in the end, this feel good Microsoft strategy ultimately fails, but hey, I’m sure there are people out there who like to scuba in their pool vs. the big bad ocean. It’s safer. Time will tell. Meanwhile, I prefer to still get lunch, and not in a box.  How about you?!  ;-)

Yours truly,

Dr. Cloud

­­­­­­­­­­­­­

Mike L. Chase, J.D., CCIE# 7226 (mchase@dincloud.com) is the EVP/Chief Technology Officer for dinCloud, a cloud service provider and transformation company that helps businesses and public/private organizations rapidly migrate to the cloud through the hosting of servers, desktops, storage, and other cloud services via its strong channel base of VARs and MSPs. Visit dinCloud on LinkedIn (News - Alert): www.linkedin.com/company/dincloud.




Edited by Stefania Viscusi




Comments powered by Disqus