Cloud // Software as a Service
Commentary
12/6/2013
01:14 PM
Reuven Cohen
Reuven Cohen
Commentary
Connect Directly
Twitter
Twitter
LinkedIn
RSS
E-Mail
50%
50%

An Inconvenient Truth About Cloud Analogies

Holy cows and puppies: Most cloud analogies make no sense, says Citrix chief cloud advocate Reuven Cohen.

We in the technology sector love analogies. We love to describe the latest trends by using historical parallels to transfer meaning from one subject to another via this linguistic crutch. It's long been the central method we've used to educate the uninformed about the benefits of a new or possibly hyped technology. The approach almost always starts with something like "The tech is like something else, only better." More often than not, it isn't anything like what it's being compared to. My proposal is simple. It's time for us to stop using analogies to market technology.

Cloud computing, which itself is an analogy or possibly a metaphor wrapped in a euphemism, has seen more than its fair share of linguistic tricks applied to it over the years. Among the more frequent is the so-called "power grid" analogy. It goes like this, "Cloud computing is like plugging into a central power grid instead of generating your own power." OK, don't like that one? There's always the thermostat analogy, "Cloud computing is like giving your IT department thermostats that they can turn on, up, down, or off based on their needs. Need extra processing power to cover your peak period? Crank it up -- and then turn it down when you're finished."

Except, it isn't.

The latest in this long line of analogies is the idea of treating your infrastructure components like cattle, opposed to pets. Recently, the concept has been promoted on various blogs and was first described by Joshua McKenty, CTO and co-founder of Piston Cloud. "The servers in today's data center are like puppies -- they've got names, and when they get sick, everything grinds to a halt while you nurse them back to health," said McKenty.

He goes on to describe his OpenStack distro, one of many, as "a system for managing your servers like cattle. You number them, and when they get sick and you have to shoot them in the head, the herd can keep moving. It takes a family of three to care for a single puppy, but a few cowboys can drive tens of thousands of cows over great distances, all while drinking whiskey."

[Want to learn more about misused cloud terms? See Cutting Through The Hype On PaaS.]

So let me get this straight. We're supposed to be able to shoot our servers in the head while drinking whiskey and still effectively manage our infrastructure? Or are we supposed to slaughter them in the datacenter for a prime cut of beef? I'm so confused.

Arguably the OpenStack community has been the worst at generating a multitude of analogies to describe the framework. Among them is the idea that OpenStack is somehow the "Linux" of the cloud-computing world. Except, again, it isn't.

In a blog post, SUSE community marketing manager Brian Proffitt explained some of the differences. "The origins of Linux are grounded in the rather organic growth of the free software Linux kernel, which was plugged together with compilers and other pieces of software to form the Linux operating system. This was a very grassroots movement, which would only later attract the interest and resources of larger corporate players," said Proffitt.

"The origins of OpenStack, which were firmly rooted in open source licenses, are very different: Also, there is really no 'core' OpenStack -- the platform is a conglomeration of tools that handle tasks like compute and storage. OpenStack has also been very heavily involved with larger corporate interests almost from the very start."

If we are to believe that projects such as OpenStack are the "Linux" of cloud, what does that make the other players like CloudStack or Eucalyptus? The BSD of cloud? Berkeley Standard Distribution has quietly become among the most used, yet least hyped open-source operating systems. Among its many uses, it forms the core of Apple's OS X operating system, making it among the most profitable, albeit indirectly, of the open-source operating systems.

Many of you will probably say that my newfound disdain for technology analogies is the pot calling the kettle black. I admit, I've been among the worst at propagating baseless analogies. Among others, I once described a musical notation application as "Google Translate for music." (Editor's note: Good analogy, Reuven.) 

I once saw someone else's analogy -- "online surge protectors" for stopping user panics online -- and adopted it: "If you think of the load balancer as an analogy for a traditional electrical surge protector, you really start to see the opportunity for cloud bursting... used in much the same way you'd use a backup power supply or even a circuit breaker."  

And for that, all I can say is, I'm sorry.

Cloud Connect Summit, March 31 – April 1 2014, offers a two-day program colocated at Interop Las Vegas developed around "10 critical cloud decisions." Cloud Connect Summit zeros in on the most pressing cloud technology, policy and organizational decisions & debates for the cloud-enabled enterprise. Cloud Connect Summit is geared towards a cross-section of disciplines with a stake in the cloud-enabled enterprise. Register for Cloud Connect Summit today.

Comment  | 
Print  | 
More Insights
Comments
Threaded  |  Newest First  |  Oldest First
Laurianne
50%
50%
Laurianne,
User Rank: Author
12/6/2013 | 1:53:12 PM
Cloud analogy fun
Good fun, Reuven. That electricity/power grid analogy has dogged the cloud community for years. Wait, I'm mixing my metaphors.
ChrisMurphy
50%
50%
ChrisMurphy,
User Rank: Author
12/6/2013 | 6:55:24 PM
Re: Cloud analogy fun
The electricity analogy was dreadful because it applied to one slice of IT -- datacenter infrastructure -- but was used by Nicholas Carr to suggest that all of the IT was doomed. As if the use of data and the use of electricity in companies were comparable, which they aren't at all.
virsingh211
50%
50%
virsingh211,
User Rank: Strategist
12/8/2013 | 1:15:42 AM
Re: Cloud analogy fun
I agree you Chris, even Gartner predicted in their report that neither electricity analogy nor water analogy works here and one top reason for same is service requirements vary widely for computing.
cbabcock
50%
50%
cbabcock,
User Rank: Strategist
12/6/2013 | 2:44:05 PM
Blame vendor copy writers, not Reuven!
Nicely said, but I think vendor product announcements with bad analogies are the real culprit, not Reuven's commentaries. I've waited years to hear a confession of wrong doing there without satisfaction. Many of the major software makers could be held accountable. They're announcements are inflated with suspect anolgies, hyperbole and multi-faceted but seamless erroneous zones. 
Lorna Garey
50%
50%
Lorna Garey,
User Rank: Author
12/6/2013 | 3:21:17 PM
You laugh, but ...
Much time, thought and discussion used to go into naming servers. Do you go with Star Trek or Star Wars characters? Maybe LoTR? Which one gets to be the Precious? It'll be a lost art. No one is going to spend time naming VMs.

However, one should be able to put a VM out of its misery without angst, so maybe the cattle analogy isn't so bad.

 
D. Henschen
50%
50%
D. Henschen,
User Rank: Author
12/6/2013 | 4:10:47 PM
You forgot the ADP analogy.
From a SaaS perspective, one analogy I've often heard used is ADP, the payroll services provider (and now an HR app services provider as well). Sometimes the ADP analogy has been used by SaaS vendors that companies have been trusting third-party providers with sensitive financial data (namely payroll) for decades. I've also heard it used by late-to-the-party SaaS vendors to show that so-called SaaS pioneers really late to the party as well (because ADP started doing this sort of thing long ago). That makes their tardiness appear less significant.

The need for this analogy is kind of fading because cloud acceptance has grown dramatically and because every vendor that counts has long since jumped on the cloud bandwagon.

 
Stratustician
100%
0%
Stratustician,
User Rank: Ninja
12/8/2013 | 4:34:21 PM
Blame it on the marketing folks
As a marketer, especially one working in the cloud space, I agree...some of these analogies make me want to drive a sharp object through my brain.  The truth is that like a lot of tech marketing, especially complicated IT concepts, marketing folks write for an audience they think isn't quite intelligent to get the concept.  Or should I say, they write for the sales folks, so they can explain it to customers.  It's a vicious cycle, which is why we have so many lovely hyped words like BYOD, Cloud, Connected Enterprise.  These have been used in every context, so the minute they come up in conversations with clients, you can see the blank stare coming on and the eyes start to roll into the back of their heads. The reality is that due tot he complexity of cloud, especially from a business transformation perspective, we (as marketers) struggle to find the balance between promoting the real benefits, with making them understandable to all the audiences.  I suppose this is what it was like when IT security started to become more mainstream!
jemison288
100%
0%
jemison288,
User Rank: Moderator
12/8/2013 | 4:53:14 PM
In Support of Cloud Analogies
Personally, I find both the cattle/puppy and electricity analogies to be profoundly helpful in understanding (a) how many organizations and individuals fail to take advantage of the cloud (while still arguing they are "on the cloud" or [as a vendor] "the cloud", and (b) how much further we have to go before we are where we want.

The cattle/puppy analogy is most helpful when demonstrating to people that their "private cloud" (read: VMware in an on-premises data center) is not taking full advantage of the cloud.  I don't think it's a particularly helpful way of distinguishing vendors (since, with enough scaffolding and the right application architectures, any virtualized environment can be a "cattle" architecture). But it absolutely helps identify, in short hand, that certain cloud deployments aren't built for failure, and they require too much individualized interaction with servers.

The electricity analogy is very helpful in identifying the compatibility and feature issues we have between clouds. The reason why the cloud isn't like electricity is because so many cloud services aren't fungible (e.g., IaaS, PaaS, SaaS--diferent APIs, different features, etc).  That said, cloud storage (S3, Google Cloud Storage, etc) actually can be like electricity, and, in fact, quite a few organizations and vendors do use it this way. We may actually be able to get to electricity for compute, but we're not there yet.
Li Tan
50%
50%
Li Tan,
User Rank: Ninja
12/9/2013 | 12:16:38 AM
Re: In Support of Cloud Analogies
None of the analogy is perfect and each of them only covers part of the truth of cloud computing. But I don't think we can get rid of it - the analogy provides very good starting point/initial view of a new technology for both IT professionals and non-technical people. We don't need analogy to be perfect since it's just a rough modelling of the real stuff. What we need is just an initial picture of new stuff in brevity. From this angle I do consider analogy not only interesting but also important.
AndrewT515
50%
50%
AndrewT515,
User Rank: Apprentice
1/5/2014 | 3:33:09 PM
Re: In Support of Cloud Analogies
@jemison288

You're absolutely right - the traditional style of server management that assumes each individual machine is highly available really does not reap the potential benefits of clouds like AWS, GCE, Rackspace and others (both on and off premises).  The key difference is often named "design for failure - in fact it's well known for an open source project whos name implies yet another analogy - Chaos Monkey (and the rest of Netflix's simian army). The interesting thing is that everyone seems to know these terms - who in our industry hasn't heard of Chaos Monkey?  Who hasn't seen how good netflix is as an always on service?

Some more academic folks like the term "recovery oriented computing" originating at Stanford.  Wether you like the cute analogies or not - this new approach to application availablility is definitely entering that first phase of the hype cycle often called hyper-growth.  Like all hypes before we should expect it to peak, hit the trough of disolusionment (when people realize it's not trivial to write applications this way) and eventually plateau.

As someone who has developed applications this way for the past five years, I'm a believer.  It works and as the folks from Netflix point out, it's a lot easier/faster than trying to reach perfection with our software.  It's cheaper/faster to assume our software is buggy as is the infrastructure. 

Frankly, what astonishes me the most is how fast this hyper-growth stage is unfolding - faster than any of the tech trends I can remember over the past 25 years.  I think it's accelerating so rapidly because it's a (analogy) bit like a rogue wave combining three hype curves together.  Whenever I come across yet another enterprise customer starting to develop livestock style applications they will almost always be developing on an open source stack ranging from open source scripting languages, frameworks, operating systems, hypervisors, cloud managers, source repositories, build tools, continuous integration tools etc.  Sure open source tools and tech like these have been around a long time, but we're seeing rapid growth particularly in enterprise development teams.  Third, these teams are almost always applying DevOps style processes usually beginning with "Continuous Delivery".  Each of these buzz terms is going through a hype cycle - but they're not entirely independent - they're very often combined together - and often to support big data, mobile, and social applications  - also going through rapid growth.  

Even if you don't like the way we're describing what's going on in our industry, you can't ignore it.  It looks like the tsunami of tech growth at least for my career.  Fortunately, I love to surf, kite surf, wind surf, etc.  Better to be on the wave than on the beach ;-)

These are my personal views/opinions and not of IBM

Andrew Trossman
8 Steps to Modern Service Management
8 Steps to Modern Service Management
ITSM as we know it is dead. SaaS helped kill it, and CIOs should be thankful. Hereís what comes next.
Register for InformationWeek Newsletters
White Papers
Current Issue
InformationWeek Tech Digest - July 22, 2014
Sophisticated attacks demand real-time risk management and continuous monitoring. Here's how federal agencies are meeting that challenge.
Flash Poll
Video
Slideshows
Twitter Feed
InformationWeek Radio
Archived InformationWeek Radio
A UBM Tech Radio episode on the changing economics of Flash storage used in data tiering -- sponsored by Dell.
Live Streaming Video
Everything You've Been Told About Mobility Is Wrong
Attend this video symposium with Sean Wisdom, Global Director of Mobility Solutions, and learn about how you can harness powerful new products to mobilize your business potential.