Home > Cloud > Should we be looking at “The Cloud”?

Should we be looking at “The Cloud”?

Recently I’ve been asked to do some cost comparisons against putting data or computer resources into the cloud.  This is of course something that many are asking about, and it’s only a good idea to check out all options.  In doing so, I realized a few things.


1) Everyone thinks the “IT Guy” – that’d be me – is concerned about his job.

This just isn’t true.  The worst that would happen, is if you’re good, you’ll go work for the big guy providing the Cloud Services.  If virtualization/data center isn’t your gig, the company almost certainly has tons of general infrastructure stuff it needs done – even if it’s optimization, automation, and standardization.   I haven’t heard of a place that has a shortage of work.

My biggest concern is budget and value.  That’s just the kind of guy I am.  If something is a better deal, or a better product, and can help the business, that’s what I want.  If it can help buy better tools or more hands, that’s (especially) what I want.  So whether the discussion is Cloud or what may be the more expensive product, if it doesn’t help or we won’t actually get to pushing half the buttons it comes with, I’d rather do without.

The reality is, while I am a “Data Center”, “Virtualization” and “Storage” guy, that’s probably 20-30% of what I do on daily basis.  Most of those things are what I do between meetings or fighting fires, or helping come up with “the rest of the plan”.  So while shipping off the IaaS might seem like a cost savings, I’m the guy who sees it most directly day to day, and that’s really not where we’re chewing up hours or dollars.


2) “Cloud” is nothing new, they’re just talking about it in relation to IT.

Here’s the thing about “The Cloud”.  If you’re looking to stand up 20 servers *tomorrow*, and/or you need them for 90-120 days, it’s a good fit.  Just like renting a car or a hotel room is a good fit for a short term.  But if you’re going long term, with the same environment, no one in their right mind would rent a car by the day or week for 3-5 years, nor would they rent a hotel room.  Yet, convert it to IT speak, and it seems like it’s a new idea and a new way to do business.

The next problem is this concept of “push a button” and “scale out”.  When the vendor or sales person is talking to the executives about this, they’re not talking about the technical side.  Sure it *can* be done.  I assume the current environment already works that way.  You just push a button, and 10-20 servers spin up, get IP’s, join the network, install all the software, configure all the settings, add the right users, update the documentation, review the change management, etc.  Because we all must have this level of automation and orchestration kicking around that we’re hiding from management and doing it the long way – we’re just doing it the hard manual way because we’re tired of having nothing to do, being so far ahead of the wave.  That must be it.


3) They’re forgetting the logistics.

Often the big push the vendor has, is that they can get their foot in the door with DEV/TEST.  They seem to forget that you need to GET your data to DEV/TEST.  If you’re testing only code, and compiling it, this can work.  If you’re testing databases, you might be doing dump and loads of TB of data daily or weekly.  You might be cloning Production VM’s to DEV/TEST.  How is all that data getting to the Cloud?  Can your bandwidth handle it?  Can you allow for the time it would take to make it happen?   How are you getting TB of data into and out of this cloud?

Is any part of what you’re testing, related to performance or capability?  If you’re using your TEST environment to figure out how long a process might take – and it doesn’t MATCH or at least closely emulate Production…. is your test going to be all that valid?


4) The vendor is in it for THEM, not for you.  They WANT you to make a mistake. 

Whether we’re talking compute and VM’s, or talking storage, you’re going to buy a certain amount.  But you’re paying for what you use, not for what you have and allowing the business to reap the benefits of having some dynamic capacity.  If you need more for a while, that’s fine – but you’re going to have to pay for it.

Does your current storage have features like deduplication or compression?  Snapshots?  Replication?  Your vendor’s may or may not.  But if it does, the benefit is going to THEM to provide them with oversubscription capability.  You’re almost certain to pay for backup as well, just like you would in your office.  Do you know if they are replicating your data offsite?  This is the part where they’re in it for them.

The mistake they want you to make?  Forgetting.  Forgetting to turn off some VM’s you don’t use.  Forgetting to delete data.  Letting your staff make 2-3 copies of the data.  This is no different than a rental car company hoping you’ll forget to fill the gas or that they bill by the mile.  Or the cell phone company that hopes you’ll check your e-mail while roaming abroad.  For things like storage, they’re going to charge based on data input.  Data *stored*, which might be reduced by deduplication, compression, or other methods, that’s to their benefit – not yours.


5) They already HAVE the Cloud.

Chances are, if you’re having this talk, you’re already virtualized.  You have a number of hosts with N+1 or better capacity.  You have growth room built into your SAN’s.  You’ve got 10GbE switching fabrics to allow for growth and density.  So from the perspective of dynamic and elasticity, you’re already there.  You may have centralized your data center in one office server room, or in a central or co-located facility.  This means that all of the data is on the other side of a WAN pipe from many of your clients or offices anyways.

So exactly how is this NOT “The Cloud”?  You access it through a straw, you can have double what you have today, and if you need to add more data, you can.


6) How do you compete?  What can you do?

Why resist?  Companies want IT to operate more as a Business to Business relationship.  They want IT to be a profit center vs a cost center.  Most IT departments don’t have a good model to do chargeback.  Most likely if they did, they tallied up their costs, then added some percentage for margin, and called it a day.  This isn’t very good marketing though.  Charge what the market will bear.  Do the cost analysis against the cloud providers.  Then undercut them.  That’s the way to take your business to market.  If you’re being told that someone can provide “Cloud Storage” for $0.10, $0.25, $0.40/GB/month…. find a number you like, and charge 20% less.  Validate that this is better than your costs.  Above all else – make a profit.  The other guy is going to.

Use the profit to continue to ensure you have excess growth capacity.  Remember that “capacity” is not always hardware and software.  You need to be sure you have hands and brains to push those buttons.  The business wants to be able to turn on a dime.  You don’t do that by having “enough to make do”, you need to be sure that you can jump.  To do this, you need to be caught up, and ready to go when the clock strikes.  You can’t be working on problems from years ago, that didn’t get taken care of then.


It’s really not all that much of a surprise.  I’ve just seen most of the pieces in the same place now.

Categories: Cloud
  1. No comments yet.
  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: