Over the years, you've probably learned that you don't talk about politics or religion at social gatherings.
If you're in the IT industry, you'd probably add "cloud" to the list of topics you approach with some caution, especially in larger settings.
I've often wondered what is it -- exactly -- that makes cloud such a divisive discussion?
We've been collectively hacking away at the cloud debate for many years now, and -- although there's been some progress in a few areas -- we're nowhere near anything that approaches an industry consensus.
James Urquhart does an admirable job of tackling this thought in his recent post "Why It's So Hard To Talk About Cloud", but -- after reading it -- I think there's something deeper going on: the incredibly wide swath of stakeholders inevitably leads to an incredibly wide range of divergent perspectives.
For those who are waiting for the debates to subside and a reasonable common ground to emerge, don't hold your breath: there's far too much at stake for all involved. Smart IT leaders will have do what all leaders usually end up doing: listening to all the perspectives, and deciding their own path forward in the face of rampant and bitter industry arguments.
And, since the potential benefits of cloud concepts are very great indeed, there's a strong incentive to plan a course of action sooner than later.
Where You Stand Depends On Where You Sit
As with any heated discussion, your personal perspectives have a lot to do with your particular vested interests. And, since cloud concepts cut across such an enormous swath of said vested interests, there are many perspectives indeed to consider.
I see people inevitably get dragged into pointless arguments around who's "right" and who's "wrong". As is the case with politics or religion, there are very few universal truths.
The most successful leaders will acknowledge the different perspectives, and chart a path forward that respects diversity of opinion without being enslaved by it.
The Business User Perspective
If you're a business user of IT, you often wonder what could be so darned hard about delivering competitive and easy-to-use IT services. Heck, we live in a world of smartphones and iPads and Google and SalesForce.com and Dropbox and all manner of easy-to-use IT environments.
What's the big problem?
You say the word "phone" and everyone thinks about small, easy-to-use devices that dramatically improve productivity.
But behind that "phone" is an wonderfully complicated and sophisticated set of network services that didn't spring into existence overnight. The people in the telecommunications industry have one perspective --but the people who buy and use their services have a decidedly different one.
To the extent that the people providing the service can adopt the perspective of the people using the service, they win.
And that's only one of the major seismic faults in the cloud discussion: the complete divergence of perspective between the people providing the IT service (the enterprise IT group) and the people using the services. The people providing the service often get frustrated because the people consuming the service have no idea how complicated and difficult this all is. And the people consuming the service point to other examples and wonder what could possibly be so darn difficult?
I did what I could to dive into an approach to bridging this particular schism with "Why IT Organizations Will Invest In Marketing".
The Business Leader Perspective
Business leaders tend to be singularly focused on making their businesses more successful. They tend to be good at it :)
When it comes to IT services, it's an increasingly important input into the business equation: much like labor or capital.
Like any other business input, you want the perfect trifecta of relatively low costs, high quality and exceptional agility. And, more than likely, you -- as a business person -- are willing to invest to get that outcome if you believe the team can deliver.
From a business leader's perspective, the interest in all things cloud is clear: is there a potentially better way to do IT? One that my existing team might not be 100% comfortable with?
If you're an experienced business leader, you're quite familiar with the phenomenon of encountering and evaluating "potentially better ways of doing things". You've done it for a few decades, maybe more. And, as a result, you've seen just about every part of your business operation transformed over time with new thinking and new approaches.
IT is certainly no exception.
You have probably come to realize that corporate functions can often become entrenched and disconnected through no fault of the good people who work there. You suspect that the IT
team might need a nudge or two to start pursuing a new line of thinking. And you -- as a business executive -- are more than willing to supply that encouragement.
If you've ever been on the receiving end of one of these executive "nudges", it can be pretty uncomfortable indeed. The natural human reaction is to assume a defensive stance: the technology isn't ready, there are security concerns, there are no industry standards, etc. etc. While all of these are reasonable statements, the experienced business executive has seen defensive reactions before, and knows how to continue to press forward.
This is a version of the "irresistible force meets immovable object" logical paradox we're all familiar with. Sometimes the immovable object wins; sometimes the irresistible force wins. Again, I've done what I could to explore this thought in my post "Learning To Compete" -- which, after all, is what most businesses really want at the end of the day.
The IT Vendor Perspective
If you sell IT products and services for a living, you have to learn to thrive in an incredibly competitive business. Industry transitions tend come through hard and fast, and you don't want to miss a big one. If you see a big wave coming, you start paddling for the wave, and perhaps figure out later exactly how you're going to ride that wave.
Cloud is one of those big waves: perhaps the biggest one we've seen in a very long time.
Even if you haven't 100% figured out what it all might mean to you as a vendor, you are immediately focused on (a) making yourself an integral part of the discussion, and (b) discrediting all others who are doing the same.
If you've been watching the Republican presidential debates in the US, you'll notice the similarities.
This intense jockeying can lead to some fairly outrageous and unproductive claims and counterclaims in the industry discussion, not to mention some rather extreme positions.
The term "cloudwashing" is too mild; perhaps "cloud campaigning" is more accurate.
If you've followed a campaign season from beginning to end, you'll notice that the discussion gets more narrow and focused as time progresses.
When it comes to the cloud campaign season, we're seeing that same narrowing starting to occur, but we're nowhere near a clear two-candidate choice.
Think about it: if every vendor is saying essentially the same thing, how do they differentiate and compete? So we have yet another important source of continuing divisiveness with regards to cloud: the inevitable tendency of vendors to jockey for position.
The good news? Many of the early primary candidates have essentially dropped out of the race, simply because they don't bring enough to the table to be relevant enough.
The IT Organizational Perspective(s)
Most IT organizations are not homogeneous entities; many appear to be collections of technology tribes that happen to be working at the same company at the same time.
The infrastructure guys have their guild, the application team has their guild, the security people, the network people, the operations staff, and so on.
In larger settings, you'll find "guilds within guilds" -- clear lines of divergent opinion, often within the same functional teams.
In an ideal world, IT professionals would prefer to orient around the needs of their company vs. preserving and enhancing the influence of their guild, but we don't live in a perfect world, do we? I have been in many, many situations where different IT stakeholders are arguing about cloud concepts between themselves, and -- yes -- there's a clear tendency to jockey for position in the new world order.
I remember a few situations where the storage guys were getting into it, and turned to me for a lifeline to support their case. Sorry, guys, I don't see it that way -- at least, not the way you're seeing it.
In my world, cloud is all about enabling IT groups to deliver competitive and attractive services that are easy to consume -- whether those services are built or brokered. The technology guilds are still important, but they have to come together in a better model that's oriented around service delivery vs. isolated components. Put differently: you're still important, but in a different way. Learn and apply the new concepts, and you'll be more valuable than ever before.
So, when it comes to a cloud discussion within an IT organizational setting, we have two important sources of friction. One is the squabbling and jockeying between the traditional technology disciplines, and a second source is the discomfort that inevitably occurs when your role is evolving faster than you are :)
The good news: there are now clear educational pathways for ambitious IT professionals to take their broad backgrounds and experiences to the next level. I shared one set of offers from EMC in "Cloud And The Act Of Being Selfish".
The IT Leader's Perspective
Many senior IT leaders tend to be rather quiet when it comes to cloud. They know something important is happening in the industry, and they're carefully evaluating their options.
Yes, there are a handful of IT leaders who have plunged in, successfully managed the transformation, and are now quite justifiably proud of what they and their teams have achieved -- but I think they are measured in the many dozens vs. many hundreds or many thousands.
Many have yet to publicly commit themselves to a course of action with regards to setting a path towards learning how to compete for their internal users, using cloud as a foundation.
For some, it's a timing issue -- everything has a season. For others, it's trying to figure out what the business really wants, which -- of course -- is made more difficult since most of their business consumers aren't quite sure what they want until they see it.
Did you really know you needed an iPad before you saw one?
For others, it's a simple people equation: do I have the support of my people and and the mandate from my peers to make the investment in a long and challenging IT transformation? And -- if the answer comes up "no" -- as it occasionally does -- what's the best alternate course of action?
Either way, most IT leaders are carefully watching the discussions, and not really participating vocally to the extent that other engaged parties are doing. As in every election, the "silent majority" will usually be the deciding factor, and "cloud" appears to be no exception.
It's Election Year In The US
We've started the multi-year process of deciding who our president will be, and a good portion of Congress as well. Although the actual election won't be until early November 2012, election fatigue has already set in for many of us.
The discussion is noisy, rancorous and occasionally outrageous. Agreements between factions are hard to come by.
There's plenty at stake as well -- from taxes to job creation to foreign policy to ... well, there's a lot on the table that essentially boils down to "who are you going to vote for?"
But, come November, we'll all march to the polls and cast our votes. If we don't, someone else will decide for us. But there will be a decision. And that will be that for the next four years.
When it comes to cloud, how will you vote?
It really depends on what cloud means to you, since so many seem to be piling on to this over hyped word. I have spent a bit over a year managing systems in the Amazon cloud and I have to say it is the most frustrating experience ever. I'm flying to Atlanta tomorrow night to start installing hardware to move my current company out of it. ROI? 6 months or less. And we get tons more functionality, uptime, less latency etc. Oh and 4 hour on site support for servers+storage.
My last company I had a similar plan to move out of EC2, and everyone was on board until the last minute, and the internal politics at the board level killed the plan despite the fact it was going to save $2.1M in the first year of operation. Within a week of the plan being killed I left the company, my co-worker left and the VP of engineering+operations left. Within a month the director of operations left. Within two months the CTO left.
I have done other cloud analysis with other providers of "Infrastructure as a Service" both for online operations as well as disaster recovery. Compared to the way an efficient shop would do it, the ROI is amazing for doing things in house. I have said for a long time competing with the cloud is like grenade fishing in a barrel. You can go buy the most expensive equipment out there and still come on top pretty easily. I had one company quote me a $3M installation charge for something I could of bought for $700k. That's the most extreme example.
Other cloud offerings mean different things to me, like SaaS like salesforce or hosted exchange or something. Those are more related to internal IT and I haven't done that in ages so I'm not sure what life is like in internal IT, other than they frequently don't have a budget to do anything and usually have to take "hand me downs" from groups like mine to run their stuff, which is unfortunate.
About two years ago now the company I was at, at the time decided to outsource exchange, they spent a good I want to say 9 months going through the process of testing and stuff and then the actual move happened. Since then I'm fairly certain they've had quite a bit more downtime than when they hosted Exchange internally. (Microsoft has had a few very public outages on that stuff in the past year).
My current boss really likes the cloud concept. I'm not as much of a fan of it myself, I can see it being useful for larger organizations where there are a lot of changes going on, but smaller companies change is less frequent (maybe we spin up 10 new servers a month at the most on average). We still need to figure out the best provisioning process in our new environment. I've been supporting Red Hat based systems for years and have stuff based around that but we're currently using Ubuntu. I've been using Debian and Ubuntu since 1998 but still much prefer Red Hat in my enterprise.
Now the way I tend to design and build stuff it is usually very very efficient so competing against public clouds is easy. I think it is not uncommon though for dolts to be running their IT and things get less efficient. The most extreme example I can remember is the Transportation Security Administration wanting to spend $7M to start a blog.
So more recently I have started using another, older term instead of cloud at least when it comes to IaaS. Which is utility computing. Utility computing and utility storage is what I am most happy. I will not ever work for a very large organization, the red tape and politics are too much for me. Utility computing from an infrastructure stand point makes a ton of sense, "cloud", less so, because the amount of logic that is involved is very complex, and most of the canned tools are not cheap.
I have been saying for a long time IaaS providers have to fix their billing model, most systems, even in the cloud use a small amount of their provisioned resources, yet you are billed for everything.
Would you be surprised if I told you our storage UTILIZATION was 5% in the Amazon cloud and that is not uncommon? Why is it so low? Because they have a cookie cutter design that gives you 1.7TB of storage when you want a virtual machine with 68GB of memory even if all you wanted was 5GB of storage. Multiply that model over and over and low single digit numbers is common. They give you 160GB for their smallest instance size. I would guesstimate my average server needs about 4GB of written storage.
CPU and memory utilization are on par with industry average being under 20% for the most part. Because EC2 provisions VMs like physical machines, you can't pool resources which is a critical failing of the design. They are working with 90s era infrastructure strategy.
When I told my boss and co-worker the utilization numbers they, like me, were not surprised(they both have more experience in EC2 than me). The fact that we are all not surprised depresses me, given the costs of this thing, we will be doing soo much better once we are out.
If your familiar with Atlanta Chuck, and know of good places to go, let me know! I don't travel around the country much this is the first flight I've taken since 2005. I normally stick to west coast road trips.
My friend from Isilon still hasn't replied to my questions I sent him weeks ago for some reason.
Have a happy holidays
Posted by: nate | December 14, 2011 at 11:34 AM
Nate:
You write the best comments by far. Unfortunately, I never have time to respond to them in detail. That doesn't mean that they're not appreciated!
I agree with the gist of your points: a well-architected and well-run IT shop has very little to worry about in regards to an external IaaS provider, even Amazon.
The catch is the "well-architected and well-run" part.
I don't get down to Atlanta all that often, but -- if I do -- I'll give you a holler. We also could connect via WebEx etc.
If you send your Isilon questions my way, I'll make sure they get answered.
-- Chuck
Posted by: Chuck Hollis | December 15, 2011 at 09:46 AM
Hello. I understand that this might not be the right place but I was not successful finding some relevant discussion.
I'd like to hear some comments and/or critics on the ADIOS Software Cloud servive developed by my company.
If somebody would be interested in playing around with it and give me some echo, I would appreciate.
It's at http://www.adioscloud.com
Thank you, Dusan.
Posted by: Dusan Daniska | January 10, 2012 at 04:25 PM