Cloud Infrastructure

10:06 AM
Randy Bias
Randy Bias
Commentary
Connect Directly
Twitter
LinkedIn
Google+
RSS
E-Mail
50%
50%

DevOps: Embrace Failure & Change

In order to benefit from cloud computing, enterprises must change their attitudes and integrate platforms like OpenStack via the DevOps model.

Agility is the widely accepted watchword for cloud computing. Customers are embracing a renaissance of thinking that focuses on the idea of DevOps as a means to achieve agility. Any discussion about DevOps, like so many buzzwords, can quickly become mired in semantic arguments, but the one truism we can all agree on is that we need to break the traditional barriers between application developers and IT infrastructure operators.

Most enterprises have created environments in which developers and operators are rewarded for fundamentally opposing goals. This disconnect has existed for as long as I can remember. Developers, operating on behalf of the lines of business, are focused almost entirely on creating new value for the business. Meanwhile, infrastructure operators are focused almost entirely on managing down risk.

For an enterprise developer, getting to market in the fastest manner possible is what matters most. Conversely, infrastructure engineers and architects focus on making sure the network, servers, and storage systems never go down.

Time after time, I hear a similar story from our customers that goes like this: "We went to the infrastructure teams with the long list of requirements for our new application. They told us it would take 18 months and $10 million before it would be ready. So we went to Amazon Web Services (AWS). We didn't get our list of infrastructure requirements and we had to change our application model, but we got to market immediately."

That's because the inherent value of AWS has less to do with its cost and more to do with its delivery model, which is on-demand, elastic, and very developer-centric. If delivering an awesome infrastructure cloud was only about providing virtual machines on demand, then Amazon, Google, and Microsoft would already have serious competition in public cloud -- and they don't. 

Developers' overriding concern for the rapid and inexpensive delivery of new applications that support new business initiatives should be obvious. Operators' need to support developers in this mode should also be obvious. If they do not, then developers will continue to go exclusively to public cloud services. Security, control, compliance, cost, and many other factors will almost always take a second seat to driving new revenue opportunities, which is as it should be for any healthy business.

Historically, operators have focused on reducing the risk of new initiatives by gold-plating the initiative: spending as much as necessary to have the "best" solutions, which of course are usually the most expensive as well. In contrast, the public cloud operators usually buy the cheapest, not the best.

Cloud applications designed via a DevOps model deploy themselves, detect server failures and adapt, and do their own data replication. Cloud-native applications route around failures, making gold-plated infrastructure superfluous. In fact, what enterprise IT infrastructure operators want to do in the new model is focus instead on building a single, enterprise-wide private cloud system that is inexpensive, easy to operate, and focused on cloud-native applications.

Now, a new trend is emerging in the enterprise. Both operators and developers realize that open-source cloud platforms such as OpenStack are the ideal building block for the new elastic cloud model. We saw this happen with Linux, which was the ideal building block for web applications and the Internet. Like Linux, OpenStack may ultimately supplant the previous proprietary incumbents that solved similar problems, while also solving whole new categories of problems.

Imagine this: Developers have an elastic cloud to target cloud-native applications; operators have built and designed this private elastic cloud to look like a public cloud; and the cloud is based on open-source tools such as OpenStack.

A wonderful thing happens because suddenly, the fear of failure evaporates. Just as with a public cloud, if a new application initiative fails, you simply shut it down -- no harm, no foul. The cost of failure is inexpensive, which fosters risk taking, which in turn creates a culture of acceptance of failed initiatives. Again, it's hard not to draw parallels to the Linux experience, where the move to industry-standard x86 servers running Linux reduced the cost of failures compared to the legacy RISC UNIX days.

It's this shift toward a new application model that embraces failure that will ultimately demolish the barrier between developers and operators in the enterprise, lead to IT harmony, and deliver true DevOps. This new model requires a change in thinking around how applications are designed and delivered, combined with a change in thinking around how private clouds are designed and delivered. Neither can stand alone.

Embrace change and failure now. Be the person in your organization who helps move to the new elastic cloud model and delivers true DevOps.

Randy is the CEO and Founder of Cloudscaling, a leader in enterprise-grade OpenStack-powered products. A long-time expert on infrastructure, he has been working on high availability scale-out systems for decades. He is widely regarded as one of the most original and ... View Full Bio
Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
Page 1 / 2   >   >>
ReturnoftheMus
50%
50%
ReturnoftheMus,
User Rank: Moderator
5/7/2014 | 8:19:17 AM
Re: One size does NOT fit ALL!
'Sure. No other verticals besides IT are being impact by the rise of mobility, social networking, and big data analytics'.

These are IT trends, which all lend themselves to enhancing customer engagement, so I'm not surprised to see these being investigated by organisations large or small.

'I said the same to a room full of ~25 CTO/CIOs for Fortune 500 companies at a Research Board meeting just a month ago and there was zero push back'.

Here I would assert IT practitioners are probably more apprehensive about their futures (me included), as the IT industry goes through another wave of transforming itself, had this been to a room full of CEOs/CFOs I'm not sure you would've got the same feedback.

'It seems self-evident to me and you are one of the first people I've heard major pushback from'.

I wouldn't describe it as 'pushback', but more of a reality check!

 

 
rbias570
50%
50%
rbias570,
User Rank: Apprentice
5/2/2014 | 8:53:46 PM
Re: This scares me...
That doesn't really make sense, does it?  The app developers objective inside the line of business is to deliver a working solution to their customers.  That's a key criteria of determining success.  Who is suggesting that has changed?

Allowing experimentation and shortening time to market for new initiatives is not a zero sum game where quality and productization efforts consequently fail.

If every new initiative inside a business has to be some gigantic world-altering 100% baked product that is instantly rolled out to every customer they have ever had with no BETA test or market validation, the failure rate would be 100%.

What if the customer is an internal customer and you are bringing up a Big Data solution using Hadoop plus a data set that already exists (e.g. Taser's live video and GPS capture at time of discharge) and providing a new way to query or correlate that data?  The internal customer may be an investigative unit (same Taser example) that does immediate on-demand investigation of discharges.  The idea being to float a trial balloon around a new product offering that can be sold to LEOs.

With cloud and DevOps you could bring this up in less than a month, vet it in another month with a couple of your closer customers, then make a go-no-go decision on whether to complete the productization process.  Essentially you bring up a BETA/ALPHA, and market test it like any good company should.

Now though, in order to do the BETA you don't need new hardware, additional procurement times, or any of the other traditional long poles that would slow you down in doing your market validation.  And, if it doesn't work, you pull the plug and that cloud capacity can be used by another.

This is already what is being done with AWS.  It's simply a question of providing an equivalent service behind the firewall for security or cost reasons.
rbias570
50%
50%
rbias570,
User Rank: Apprentice
5/2/2014 | 8:42:38 PM
Re: One size does NOT fit ALL!
Sure.  No other verticals besides IT are being impact by the rise of mobility, social networking, and big data analytics.

Hyatt isn't hiring cloud app developers by the boat load. Fidelity hasn't radically changed their internal development model.  Retail businesses aren't scrambling to become IT businesses by acquiring tech companies (http://techcrunch.com/2013/10/02/staples-buys-runa-to-square-up-to-amazon-in-the-e-commerce-game-for-office-supplies/).  The retail vertical isn't opening up R&D shops to invest into IT:

http://www.stapleslabs.com

http://www.walmartlabs.com

The "Other" revenue line for AWS is the fastest growing segment of their business (58% y/y compared to 8% for their original business line).  AWS is not "profitable" because Amazon plows all of their revenue back into the business.

We're going to have to agree to disagree I guess.  It seems self-evident to me and you are one of the first people I've heard major pushback from.  I said the same to a room full of ~25 CTO/CIOs for Fortune 500 companies at a Research Board meeting just a month ago and there was zero push back.  And it wasn't like this was a group that was shy to push back, they simply agreed with me.

This isn't FUD, it's reality.
ReturnoftheMus
50%
50%
ReturnoftheMus,
User Rank: Moderator
5/2/2014 | 5:13:01 AM
Re: One size does NOT fit ALL!
I hate to do it to you, but going to have to dissect your post, hopefully alleviating the crossed-wires we keep getting.

'That doesn't really make sense if you think about it'.

Here I think we agree? Small and Medium companies are prime targets for cloud-based services, they tend not to have the inhouse resources or expertise to develop their own applications and have typically been buying them off-the-shelf so to speak, the vast majority of the staff in these types of organisation are customer facing and customer-focused with very few back office staff.

'Amazon Web Services incredible growth is largely because of small and mid-market enterprises realizing that they can now experiment with relative impunity'.

Again here we are agreed, but let's not forget that the bulk of Amazon's $74bln in annual revenues is not derived from AWS, in fact I'm not even sure if AWS is yet a profit making venture?

'Netflix is a midmarket enterprise for example. Also if you look at the latest RightScale State of the Cloud 2014 report you'll see that SMBs and mid-market are adopting technologies like Google App Engine and Google Compute Engine at a very high rate'.

Netflix is no more than a classic business success story, it wasn't the first and won't be the last. Yes, SMB's and the midmarket are adopting technologies like Google App Engine and Google Compute Engine, the same way they've adopted SalesForce and Force.com, the same way they're adopting Office365 and Microsoft Azure, the same way they'll adopt HP, IBM and Oracle cloud-based offerings. In many cases the same software and applications to which they're accustomed, only consumed in a different manner.

'The point is that business agility is now an imperative'

When was it never? Although one could argue how imperative very much depends on your industry sector, you need look no further than how crowed and congested the cloud space is becoming, then contrast that with a highly regulated industries like defence and healthcare. 

'Most businesses are beginning to understand that they are technology businesses first and their particular vertical second'.

Probably, the worst case of FUD coming out of the IT industry since the dot.com bubble.  Let's be frank, the only industry in a major transition right now is the IT industry itself.
rjones2818
50%
50%
rjones2818,
User Rank: Apprentice
5/1/2014 | 1:47:46 PM
This scares me...
"Developers' overriding concern for the rapid and inexpensive delivery of new applications that support new business initiatives should be obvious. Operators' need to support developers in this mode should also be obvious. If they do not, then developers will continue to go exclusively to public cloud services. Security, control, compliance, cost, and many other factors will almost always take a second seat to driving new revenue opportunities, which is as it should be for any healthy business."

 

Where in the name of whatever you find holy (money, I suppose) do you fit customers into this worldview?  The argument will be that the customer will benefit from all of this, as they may.  The problem is that the experimentation, etc., will be being don on the customer's time and on their dime even more so than before.  Are they being sold that their product/app/whatever isn't going to work all the time like it should?


The customer comes first, not the business.  I'm sure there are good reasons to have devops as part of your bag of tricks.  Relying on it and agility leaves open the door of not supplying the customer with what could be considered a finished product (or even a working one).
rbias570
50%
50%
rbias570,
User Rank: Apprentice
4/30/2014 | 12:50:11 PM
Re: One size does NOT fit ALL!
That doesn't really make sense if you think about it.  Amazon Web Services incredible growth is largely because of small and mid-market enterprises realizing that they can now experiment with relative impunity.  Netflix is a midmarket enterprise for example.  Also if you look at the latest RightScale State of the Cloud 2014 report you'll see that SMBs and mid-market are adopting technologies like Google App Engine and Google Compute Engine at a very high rate.

Craig Le Clair of Forrester said it best:

"Seventy percent of the companies that were on the Fortune 1000 list a mere 10 years ago have now vanished – unable to adapt to change ..."

Original here: http://blogs.forrester.com/craig_le_clair/13-09-09-make_business_agility_a_key_corporate_attribute_it_could_be_what_saves_you

The point is that business agility is now an imperative.  Most businesses are beginning to understand that they are technology businesses first and their particular vertical second.  Very traditional businesses, like Hyatt, are adopting DevOps, and bringing about the Netflix-ification of their business with FULL support from the C-level execs:

http://www.slideshare.net/mokeefe/devops-unicorns

This is not a blip, but ultimately, a sea change in how enterprises operate.  You can be part of the solution and help it manifest or part of the old guard that resists, and get run over.

It's really that simple.
ReturnoftheMus
50%
50%
ReturnoftheMus,
User Rank: Moderator
4/30/2014 | 6:09:04 AM
One size does NOT fit ALL!
As an advocate of the cloud there's not an awful lot that has been said that doesn't resonate, however I do think the rhetoric coming out of the IT industry boarders on eccentricity. Contrary to popular belief, not every company is an IT company and nor do they want to be.

Start-ups generally have the luxury of being able to start with a blank canvas (with no guarantee of success), however well funded or not as the case may be.

Large enterprises tend to have much deeper pockets than most and constantly assess risk, whether through SWOT analysis or other mechanisms and are generally not shy in making investments, whether through internally run projects or use of external vehicles.

In the non-tech mid-market for many failure isn't really an option,  they can ill-afford to experiment, they don't have the infrastructure or resources and are reliant on the tech industry to come up with solutions, here we are already seeing shadow IT take hold and very much part of the mix.

Hence, the need to adopt your own so-called 'DevOps' model is fairly immaterial when utilising services from an Amazon, Google or Microsoft who I'm quite sure are more than  happy to extend their services to on-premise, minus the failures.

Other than that I very much look forward to your next 'State of the Stack' update.
Susan Fogarty
50%
50%
Susan Fogarty,
User Rank: Strategist
4/29/2014 | 9:27:49 PM
Re: We need to find what works through opportunity to fail
Charlie, your points about continuous delivery and survivable failure are very important in this discussion, as is Randy's statement about reducing the cost to fail. Along with a change in attitude you also need a change in process that allows that failure much earlier so that the business can adapt and try again (that's where failing faster comes in). I think that is what makes the difference.
Charlie Babcock
100%
0%
Charlie Babcock,
User Rank: Apprentice
4/29/2014 | 8:05:18 PM
We need to find what works through opportunity to fail
Randy isn't talking about failure in the traditional sense of organizing big waterfall projects, then having them deliver late, overbudget or not at all. Or having a software update fail -- again. Rather, he means the opportunity to experiment within the business, not knowing whether the experiment will succeed or fail. By using the word DevOps, we fuse two seemingly opposites together, which at  least captures the notion of unity but otherwise leaves people puzzled. If you adopt the notion of continuous delivery, where one application after another can be launched into production, or one application can be frequently modified, then you have the environment for business experiment and survivable failure.
rbias570
50%
50%
rbias570,
User Rank: Apprentice
4/29/2014 | 1:10:21 PM
Re: Are you sure this sends the right message?
Startups are famous for a "fail faster" mentality (http://www.thedailybeast.com/articles/2014/01/05/fail-fast-fail-often-how-losing-can-help-you-win.html).

Enterprises need to operate at startup speed or, at least, the various departments and groups need to be empowered to do so.  If they are not, then those groups will simply go to Amazon Web Services and shadow IT will take hold.

I understand that there is a culture inside that impugns failure, but this is rapidly changing as cloud and DevOps have become a forcing function.  Failure isn't just important, it's an imperative.

How do we learn without failure?  How do we advance?

The reality is that failure is not the problem, but the cost of failure.  As the cost of failure reduces our corresponding fear of it should reduce as well.

Enter cloud and DevOps.
Page 1 / 2   >   >>
Cartoon
Slideshows
Audio Interviews
Archived Audio Interviews
Jeremy Schulman, founder of Schprockits, a network automation startup operating in stealth mode, joins us to explore whether networking professionals all need to learn programming in order to remain employed.
White Papers
Register for Network Computing Newsletters
Current Issue
2014 Private Cloud Survey
2014 Private Cloud Survey
Respondents are on a roll: 53% brought their private clouds from concept to production in less than one year, and 60% ­extend their clouds across multiple datacenters. But expertise is scarce, with 51% saying acquiring skilled employees is a roadblock.
Video
Twitter Feed