308220

308220

Debates are fun, especially when you are jousting with such fantastic and respected cloud dignitaries as Chris Hoff, Adrian Cockcroft and Simon Wardley.

Come on, I mean who wouldn’t enjoy a good natured, well intended, yet fierce back and forth about the various clouds and their various philosophies ? Well, March 12 was a day to remember…a debate of almost biblical proportion erupted on twitter, ignited largely by my experience of listening intently to Cockcroft at a lunch meeting a day earlier and then declaring the end of my presidency of the private cloud (more on that later) whilst Hoff, in his usual inimitable style, tried a double ankle grab sweep (remotely, of course) to get me to confess that I am essentially a fraud and a disbeliever in my own prior support of all things private cloud. To add a little spice, the ever-pragmatic Simon Wardley weighed in with a collection of excellent comments, many of which are so blindingly simple in concept that they represent and are testament to his wealth of knowledge and deep understanding in these areas.

Firstly, I want to make a couple of important points:

1) Although the use of the word “debate” in and of itself may be challenged, the exchange was fun, passionate and heated, yet very difficult to execute on twitter (for obvious reasons) without certain comments having the potential to be misinterpreted as personal, rather than general. I have, at all times, the utmost respect for anyone who cares enough and feels strongly about a subject or viewpoint to offer it up in a public forum and especially for those with whom I was engaged.

2) I am no analyst, expert or cloud executive. I am simply a customer, or potential customer, of “the cloud” and the technologies and vendors that provide it, regardless of denomination. Therefore, everything I opine on and the definitions I use are in “context” of my experience in my day job and not with any purist slant or bias.

So, with that out of the way, let’s go back to Friday, March 11th. Location : Netflix HQ, Los Gatos, CA and a lunchtime appointment with Adrian Cockcroft. I was there to chat to a fellow Brit, one who has a great success story in the public cloud, and with whom we share a mutual vendor in a particular technology area. I was there to listen and learn from someone who has, essentially, done the same as we have done but at the other end of the cloud spectrum. We did private. He did public. Game on.

We are almost 100% in the public cloud. We don’t really have IT ops people as the Devs run the platform. Everything we do is in Java

Those three comments were pretty much all it needed to galvanize my feeling that I was not in Kansas anymore. This wasn’t a traditional enterprise. This was an organization who built their platform and their digital business for the cloud and as the story unfolded, including discussions around how the organization was structured, what the culture was like and which technologies are employed, I realized that, to a certain extent, this is how and where I would like to drive our organization. I also realized that we could literally be light years away.

The subsequent resignation of my self imposed title of President of The Private Cloud was really nothing more than a frustrated exhalation of four years of hard work (yes, it took us that long to build our private cloud) and a shuddering and uncomfortable acceptance that although we have done much, much more than other organizations, the road to where Netflix are today (one which I believe will be traveled by many organizations over the next few years) even with the head start we have, is, as The Beatles said, a long and winding one.

Taking a trip down memory lane, let me explain that we built our private cloud as a result of business drivers (and alignment) for two main reasons:

1) Speed of delivery of services (critical to our core business)

2) Geographical reach (we are working in more locations than ever before)

There is no doubt that our effort, in “context”, has been incredibly successful in terms of achieving a marked difference to 1) and 2) above. We can now be more responsive than ever before and we are no longer “IT the problem”. Is it how I would ultimately want it ? No, of course not, but when you spend a lot of blood, sweat and tears building something, you learn to love it. It’s human nature to defend it. That doesn’t mean it’s your ultimate personal goal or preference, but it is the result of the best you can do with the constraints you have, but in the same way you would still love your kids if they were ugly, you’re kinda stuck with what you have.

Today, in our environment, there has been very little work done in making material architectural changes to LoB applications, such that they can be deployed, at scale, on public cloud infrastructures. Unlike Netflix. I am not convinced that LoB owners can see any value in change. There is an air of “if it ain’t broke…” and when they own the $$, this is an area that is, and will continue to be a massive challenge for us and other organizations too. Clearly, there are major differences between what we (and other enterprises) do and what Netflix do. Their app is their core business. Our apps support our core business. That nuance can not be lost in trying to make sense of the differences and this juxtaposition is why I made the comment, in context, that I will staunchly defend:

Building the private cloud that is devoid of any plan or funding to make architectural changes to today’s enterprise applications does not provide us any tangible transitional advantage, nor does it position our organization to make a move to public cloud

Let me qualify that by adding the following:

Unless that transition or move includes utilizing services such as Amazon VPC or CloudSwitch to provide overlay networking capabilities, logically extending our environment into a physically separate managed infrastructure, which by our definition and context would represent a hybrid cloud solution

Hoff suggested that my defintion of hybrid cloud was limited to “overlay networking” (due to a comment I made which was hard to contextualize) and to clear this up, I clearly understand that there are hybrid cloud solutions that can operate easily without “VPN” technologies, just not in our world because the current architectures and technologies employed would render it almost impossible to do so. Wardley calls this “an enterprise cloud”. That was a nomenclature I hadn’t heard before, but it sort of makes sense, even though I don’t think Simon subscribes to it himself.

I have written previously about “opportunity” in which I explore a real world example of the relatively small amounts of funding available to make major changes in the LoB space. Interestingly, Cockcroft’s recent (and quite brilliant) blog post challenges the very heart of these issues, claiming:

There is no technical reason for private cloud to exist, it’s all $, FUD and internal politics.

I find it hard to disagree with this as a philosophy (although I could point to policies being a missing part of the equation and certainly debate the rest of the content ad nauseam) and I am pretty sure that my vociferous concurrence with this acted as the blue touch paper for the fireworks that ensured, but, I will reinforce one point, remembering the point I made earlier about “transitional positioning”:

Today, there is every reason for private cloud to exist because it some cases, like in ours where $ ownership for infrastructure and LoB applications lives in different fiefdoms, it can bring efficiencies and value in areas where you can absolutely NOT get the stakeholder alignment and buy in that you need to deal with the $, FUD and internal politics that are barriers to public cloud

We will never be Netflix. I accept that. I also accept that there are uses cases galore on all sides of the debate and I believe that pigeon holing any solution or approach is simply bad form. However, I firmly believe that you can never learn enough about what others are doing and how they are doing it. Open discussions and exchanges of information with people who are willing, knowledgeable, experienced thought leaders can only be good for ourselves, our ecosystem and will help contribute to death of the thought that the “one size fits all” is a viable cloud model. That funeral can’t come soon enough.

Today, our private cloud is my ugly kid. I made it and therefore I love it. That doesn’t mean that I don’t want my next kid to have Hollywood good looks. One day. Not everyone makes it to Hollywood. Ask those who auditioned on American Idol.

So, there it is. I am insanely jealous of what and how Cockcroft and his team have done, but jealous in a good way. I am also fortunate that our businesses do not conflict in any way, shape or form, so that I can continue to tap in to his knowledge and share information reciprocally.

There is no right or wrong, there is just opportunity and timing. To his credit, Hoff nailed it at the end of our discussion:

Use the right tool for the right job and the right time and the right cost

That might just be the most sensible strategy for all of us.

Leave a Reply