Selling Cloud to Your Boss…and Yourself!

By on 04/08/2011.

It’s really funny to me how most conversations around Cloud often times circle back to Virtualization 101.  And most times, it’s current VMware admins asking these sorts of questions about vCloud, still struggling to understand fundamentals.  So I want to try to help.

Why do we need it?
What will be different about our environment?
What hardware should we use [for Cloud]?
Why do we need shared storage?
Which storage vendor?
Which server vendor for hosts?
Should we use blades or rack servers?
Will my applications run just as well in this [Cloud] architecture?

Any of these questions sound familiar to anyone else?  Weren’t we asking/answering these same questions back in 2007 and ahead about why customers should virtualize in the first place, as VMware and consolidation began to take hold?

Well, there is a reason.  It’s my opinion that we’re on the precipice of the next major paradigm shift in the way datacenters will be run.

Before we get into the details of said shift, let’s take a trip down memory lane of the last 5 or so years.

First, we abstracted the servers from the apps and underlying operating systems to see what this VMware stuff was all about, most reasonably seen in Test/Dev environments.

Next, we began to take seriously the concept of running production workloads in a virtual environment as the product(s) matured and admins became comfortable working in this environment. This was your low-hanging fruit, “junk VMs,” and simple servers, such as web servers.

Then we began to figure out the real benefit:  How it’s no longer about server/disks for each request/app/dept, but how we could begin to pool the resources and standardize deployments in shared infrastructures, and build an “on-demand resources” environment.

And now, we hit the most popular “OK, now what do we do?” stumbling block.  You’ve consolidated your datacenter on your favorite flavor of shared storage.  If you’re a NetApp customer, you’re likely seeing huge returns in storage space savings from block-level deduplication, and you’re likely taking advantage of our snapshot-based backup architectures to leverage near-instant backup and recovery mechanisms, whether it’s our SnapManager products for applications and virtual infrastructures, or mirroring with SnapMirror, or long-term retention snapshots on off-site storage using SnapVault.

Here’s where I’d like to begin to fill in the blanks.  I’d first like you to take the time to go and read these posts by Scott Baker, and his posts around Service-Oriented Architectures.

“The Six Million Dollar” Service Oriented Architecture

Currently, he has published three of the four parts of the series, and I’m looking forward to reading the fourth myself.

Why am I bringing this up?  Messaging.

No, I’m not talking about Exchange, or e-mail in general.

To me, there is a fundamental likeness in message between what we’ve done for the past few years, and what we’re going to be doing in the next few years.  At the end of the day, there is not much difference in the internal private cloud you’ve likely built today and what will be used to host your stuff in the future.  Studies are showing that the three big players in the market are VMware, Microsoft, and Citrix, in that order.  All are making strides to insure your experience in transitioning to “the next step” are easy.  This is done with management tooling.  We’ll get to that shortly…

 

People, Process, & Politics

When we begin to talk about the next steps, people are quick to question the technology, the hardware, the software.  But those aren’t really the problem, are they?  Those three big bolded words above are.  How do we, as vendors, help you, as customers, drive this message home to your powers-that-be?  Well, it’s simple really.  How did you convince everyone to go along with consolidating your servers and storage in the first place?

BINGO.

This is what I’m getting at.  The messaging, and the way it is delivered needs to be handled in the same fashion that it was handled back in the “old days.” Moving to ITaaS is a BIG DEAL.  It’s a fundamental change in philosophy of how IT departments are run.  You have to institute new procedures to handle the accounting.  You have to institute new methodologies into provisioning, backup, and regulatory compliance.  For those of you that answer to SOX/HIPAA/etc, there will need to be updates to that documentation and sent off for approval.

Point is: THESE ARE ALL THINGS WE’VE DONE BEFORE.  IT’S NOTHING TO FEAR!

So how do we finish off the diagram from above?  Management.  Most shops, once consolidated, have no idea how to manage all the many layers of virtualization, network, storage, applications, etc.  Powershell emerged with a vengeance during this time, and we began to see a universal abstraction of interaction at all layers via Powershell.  ISV’s began to build applications linked to underlying API’s to automate and display/report things into Single Pane’s of Glass.  Problem is, we needed a single pane of glass for all the different single pane’s of glass.  Ha!

The idea of the fully-automated, self-managing datacenter should be your next goal once you’re consolidated.  If you’re not 100% virtualized, get there. Whatever it takes, and before you do anything else.  I have numerous articles and references pointing to the idea that virtualization is no longer the bottleneck, and if you’d like some of them, respond below or hit me up on twitter and I’ll get you what you need.

Aside: Actually, the funny part is, working for a storage vendor now, I believe we’ve come full circle to where we were 10-12 years ago:  The spinning disk.  Which is why you see all of them scrambling to offer you SSD’s, disk-tiering, etc.   Thankfully, here at NetApp, we don’t have too much to worry about, because combined with technologies in our WAFL, NVRAM, and FlashCache architectures, we alleviate those problems for you without having to buy extra spindles.

Now then…

You’ve consolidated your servers onto VMware and shared storage arrays?  CHECK.
You automated your datacenter, and updated your documentation to reflect as such?  CHECK.

Now what?

Now it’s time to start looking at the bigger picture:  Abstracting the datacenter itself.  By abstracting the datacenter, you begin to fully implement the idea of Service-Oriented Architectures, SLA’s, and pools of compute.  And you make this self-service to your internal departments, or to your customers.

What I want you to understand is this:

Whether you’re a Service Provider servicing customers, or an internal IT shop servicing departments, the principles are one in the same.

ITaaS = ITaaS = ITaaS.

Let NetApp help you get here.  With our tight integrations to VMware’s vCloud Director, our Storage Cataloguing capabilities in OnCommand, our vmoapp storage provisioning plugin for vCenter Orchestrator, REAL deduplication, snapshot backup & recovery, we are enabling the movement to the Cloud.

This is the future. And it’s not far off, folks…

As always, thanks for reading!  Please feel free to continue the discussion in the comments below! And I look forward to any feedback or followup you’d like to have!

-Nick

No Comments