Topics


Blogs


Forums


Samples


Media


Labs


Resources

 




DevCentral > Weblogs > Lori MacVittie - Two Different Socks
 No soup for you!
posted on Wednesday, April 22, 2009 3:18 AM

Automation isn’t some special brand of soup and there’s no “automation nazi” who can deny access to its benefits.

noautomationThe recent McKinsey report on cloud computing has pundits everywhere choking on their donuts and scrambling to dispute the report’s findings, which essentially end up saying “cloud ain’t cheaper.”

I’m not going to rehash the arguments. I’m not going to analyze the report. But I am going to dig into a few comments on the report by Thorsten at RightScale who started off by saying:

Its claim that cloud computing (in the guise of EC2) ends up being more expensive per server month for large enterprises than doing it in-house seems fatally flawed. In particular, it doesn’t seem to be accounting for the costs correctly and it completely fails overlook the benefits of automation in the cloud which ultimately leads to a revolution in the way compute resources are consumed.” [emphasis added]

On the one hand, he’s probably right. Most cost-based analysis of software, hardware, outsourcing, etc… tend to fail to recognize soft-costs associated with implementation. So the reality is that the McKinsey report may indeed be flawed in that the cost-reductions associated with automation and thus the streamlining of processes were not be considered.

And yet the statement appears to imply that automation is peculiar to “the cloud”, at least as it is perceived by purists who believe such constructs can only exist “out there” and cannot exist “in a localized, on-premise cloud” environment. That’s not necessarily what Thorsten was trying to say, but there is a feeling of exclusivity often exuded by some – usually providers - that attributes all the benefits of automation and virtualization to “public” cloud environments alone.  But there is not just one provider of automation, and its existence in the cloud does not preclude its use in other environments, including the data center. Automation isn’t some special brand of soup, and there’s no “automation nazi” who can deny access to its tasty benefits.

This ultimately means that some of the arguments against the McKinsey report are equally flawed.


AUTOMATION ISN’T JUST FOR CLOUD PROVIDERS


Thorsten later argues a bit more on costs associated with automation:

What the report seems to completely overlook is the possible reduction in sysadmin costs. One of the huge benefits of the cloud is that the entire computing infrastructure can be automated. Top to bottom. That saves a lot of sysadmin labor and in the end it means that requisitioning more compute capacity can be done by the end user somewhere in a business unit instead of being an IT chore.

If it can be automated “out there” it can be automated “in here”. Surprise! That will save a lot of sysadmin labor inside the data center, too.

Granted, the administrative costs (operating expenses) should be reduced even further by using the “public cloud” simply because those costs, like resources, are shared. Which is probably the point Thorsten is trying to make, but that fails to really come through. Maintaining a data center presence necessitates management and maintenance which will continue to incur operating expenses in a local implementation that obviously cannot be shared with other organizations. The cost reduction from automation of processes, however, should be the same regardless of where “the cloud” physically resides.

The benefits of automation in general are not peculiar to cloud computing providers, it’s simply that the cloud computing providers arrived on the scene first. In fact, the need for automation in the organizational data center as a means of reducing costs is certain to drive innovation in existing network and systems’ management vendors as well as spur the creation of startups providing “cloud computing automation” software designed to minimize human intervention and automate all those mundane tasks that are driving the costs of IT through the roof.

It isn’t like administrators haven’t spent years automating tedious processes in the first place whether via custom scripts or command-line tools or full-blown automation systems. They “get” it, and they understand that automation results in a huge savings in terms of time and money. They certainly, therefore, are capable of applying that same logic – and expertise - to on-premise cloud computing and virtualized environments.

Interestingly enough, automation of an on-premise internal cloud would free up IT resources, such that another of Thorsten’s arguments might also become moot.

The report also doesn’t take into account the cost of the red tape that surrounds corporate IT. Things the business can’t do because IT can’t support them.

If IT isn’t spending its time worrying about all the little mundane administrative tasks then it just might have time to support the business in the manner the business would like to become accustomed.


COST IS ONLY ONE PART OF THE EQUATION

There are so many other factors that go into the decision to use or not use the cloud that basing a decision “to cloud or not to cloud” solely on alasmydatacenter financial factors is just not reasonable.

A recent ZDNet story cited a silicon.com study in which CIOs overwhelming decided that cloud was not a viable solution for complete outsourcing at this time for several reasons, among them security, reliability, and maturity. There are also issues regarding control and the ability of the business and IT to influence and manage their applications in the way they have become accustomed when moving them to a public cloud.

So while cost is certainly one of the most influential factors in the decision-making equation “to cloud” or “not to cloud”, it isn’t the only one. And while the McKinsey report may be ignoring some of the cost issues that are certainly important to consider, some of the outrage directed at the McKinsey report is equally guilty of ignoring all the other factors that go into such decisions.

Follow me on Twitter View Lori's profile on SlideShare friendfeedicon_facebook AddThis Feed Button Bookmark and Share

Related blogs & articles:



 
      

Feedback


4/22/2009 5:34 AM
Gravatar Great article Lori.

Automation _is_ the key for managing scalability. This is true for physical and virtualized environments, as well as for deskside, on-premise datacenter or cloud environments.

But clouds not only bring in scalabilitiy, but also more complexity for the admins by adding a new abstraction layer, like virtualization did before. And here comes the caveat: Common Automation solutions are still (nice and comfortable) Admin-Tools requiring the user to bear the whole computing environment in mind for configuring "Automations", "Rules", "Actions" and "Correlations".

From my point of view - working on smart Automation solutions - only true dynamic, intelligent Automation will do the job. Clouds are beyond human admins perception and this is the reason why there is so much discourse on this topic.
Roland Judas

4/22/2009 6:15 AM
Gravatar Good article. I think one of the reasons automation isn't used more in the data center is that it generally isn't uniform enough to really benefit from touchless automation. Version differences between system software can really muck things up and that is one of the reasons commercial automation systems cost so much to purchase and maintain--someone has to keep the scripts current from version to version.

I'd think that in a cloud environment, you'd want to have a highly uniform data center. That would greatly simplify the abstraction layer Roland mentions and bring it within the realm of more IT shops.
mike fratto

4/22/2009 6:20 AM
Gravatar The main attraction for clouds is their elasticity. If we take electrical grid for example. Nuclear power plants are probably one of the cleanest and cheapest ways to get electricity, but you can not turn them on/off on drop of a dime. They produce constant huge amount of power. This serves as the base in many countries. However where does one get power when for a short period of time one is short of it? You can quickly spin gas power plant (they do turn on/off pretty quick) but you have to have one ready for it and it is not cheap and doesn't do much when it is off. Or you can do reverse water plant (when one pumps water up during low demand powering by other sources, such as nuclear power plant electricity) but you still have to have one of those handy. Or alternatively you can order some of the energy from other power grids/countries at given point that have such energy available to sell to you.

Per MWh such power is very expensive compared to nuclear power plant you own, but how much does constructing spare power plants cost so you can run it for a total a month per year?
Izzy

4/22/2009 10:25 AM
Gravatar Nice article, and I agree with the points on cloud not having a "lock" on automation.

Wrt the McKinsey report, though, I think one of the main oversights is that automated, elastic clouds can alleviate the need to overprovision. So it's not the automation itself, but the burst/overflow case it helps enable.
Glenn Dasmalchi
 Leave Feedback
Title  
Name  
Email
Url
Comments   
Please add 3 and 5 and type the answer here: