Monday, July 12, 2010

Should You Build or Buy Your Cloud?

It's the age old question in IT, the question of whether or not to build it yourself or just buy it off the shelf. Lately, I seem to be hearing the questions again and again. It seems that for some reason some IT guys have gotten it into their head that if they adopt a cloud infrastructure platform, either hosted or in house, they're going to lose their jobs. So the only choice is to build it. I think the reasoning is if you build it, you will control it, and your company will have no choice but to keep you around. Unfortunately the answer isn't so cut and dry.

The rationale for building it yourself has been around as long as IT. There have always been various reasons for it, from there weren't any systems that could delivery what we needed, or we're different, we're smarter, we're bigger.. you get the point.

The real question you need to ask yourself is where does youe strengths as an organization lay? As a software developer or selling some other core business? For most it's the latter. Building your own cloud software is fraught with risk. One such example is a major hosting firm who spent 16 months building their own cloud IaaS platform only to realize that the assumptions they made about the potential cloud market opportunity had changed and their platform couldn't deliver the technical requirements of their new customer reality. More to the point, their platform wasn't what their targeted customers wanted to buy. Compounding their problem was the platform they built themselves didn't actually work - period. The key system engineer left mid-way through the project, forcing the company to find a replacement as well as inducing a major delay in the development. Additionally poor documentation meant those replacements had no practical way to continue what had been started previously. Needless to say, several million dollars later the project did launch, only to be promptly replaced a by a turn key IaaS platform.

Then there is the question of service differentiation to which I say, if you choose an extensible cloud platform, then you're able to differentiate faster than you could if you build it yourself. Business is about adapting to market conditions. Building it yourself mean longer development cycles and potentially less adaptability. Customizing an existing platform, one that provides you a template for success and best practices is inherently less risky or time consuming. The real question you should be asking is can I deploy this cloud platform in a way that allows my business to be unique? If the answer is no, than find a platform that can. If you still can't find one, then build it yourself. But be prepared that you should now consider yourself a software developer.

As the founder of a IaaS platform vendor, I freely admit I am bias toward buying a platform over building it yourself. My reasoning is simple, our business is building IaaS platforms for service providers. Is it yours? If you answered no, my comment is unless you plan to get into the software business, building it yourself will only serve to add un-need risk, uncertainty and potential failure to your IT operations. Something I think we can all agree you should avoid.

Sunday, July 11, 2010

Embracing Your Niche in The Cloud

With all the talk and the hype surrounding cloud computing many seem to be missing a major factor -- both in terms of the growth potential as well as the current opportunity for cloud computing products and services. Although the tech media and analysts love to tell you the cloud is everything including a $160 billion plus opportunity, like it or not, Cloud Computing is still an emerging niche market and exists only as part of a much larger market segment. And I'm here to tell you that as soon as you start embracing this fact, the sooner you will start to capitalize on the opportunity.

Regardless of your industry or market segment, every single product or service that is sold today can be defined by its market niche. Of course there are the products aimed at wider demographic audiences otherwise known as mainstream niches. But those markets tend to take years or even decades to mature. As an example think of the broader web hosting industry compared to that of the shared hosting, VPS, CDN, or the managed / dedicated server markets. Those companies that arguably have had the most success in each of these markets focused on winning in their particular niches. Rackspace within managed hosting sector or Akamai in the CDN space. Both can be considered as part of the broader hosting markets, but both have significant differentiation and more importantly success with in their particular niches. Both have also been able to charge signifcantly more than the previous generations of services within the broader hosting market.

Also being first in market doesn't necessarily mean you're going to own it. When looking at niche markets it's interesting point out that narrower demographics (PaaS, SaaS or IaaS in contrast to VPS hosting) lead to elevated prices due to the concept called the price elasticity of demand. (Which is a good thing) In other words, a niche market is a highly specialized market that allows you to survive among the competition from numerous much larger and broader focused competitors.

As a further example think of Amazon (the book version of course) versus your local community corner book store. The economics that define success for that corner book store are significantly different than what Amazon would consider a success both from a profit margin as well as a volume point of view. The corner book store wins by effectively differentiating itself from it's much larger competitor. It's books are more expensive to buy, but possibly you can also buy a coffee or browse actual physical books or even have a conversation with a human. This differentiation attracts a unique customer profile and caters to an alternative market segment. A segment that possibly is willing to spend more for something they could have gotten cheaper at Amazon. These differentiators allow the corner book store to compete and even win business (within their niche) from that much larger competitor even though the price is higher. The same concept applies to cloud computing.

I suppose what I am saying is you will never win if your goal is be Amazon. You will win by not being Amazon. By being different. By embracing your niche.

Friday, July 9, 2010

Do Customers Really Care About Cloud API's?

Interesting post by Ellen Rubin of CloudSwitch asking if Amazon is the Official Cloud Standard? Her post was inspired by a claim that Amazon’s API should be the basis for an industry standard. Something I've long been against for the simple reason that choice / innovation is good for business. Although I agree with Ellen that AWS has made huge contributions to advance cloud computing. And also agree that "their API is highly proven and widely used, their cloud is highly scalable, and they have by far the biggest traction of any cloud". But the question I ask is do cloud customers really care about the API, so much as the applications and sevice levels applied higher up the stack?

At Enomaly we currently have customers launching clouds around the globe, each of which have their own feature requests ranging from various storage approaches to any number of unique technical requirements. Out of all the requests we hear on a daily basis, the Amazon API is almost never is requested. Those who do request it are typically in the government or academic spaces. When it is, it's typically part of a broader RFP where it's mostly a check box and part of a laundry list of requirements. When pushed the answer is typically, -- it's not important. So I ask why the fascination with the AWS API's as a sales pitch when it appears neither service providers or their end customer really care? More to the point, why aren't there any other major cloud providers who support the format other than Amazon? The VMware API or even the Enomaly API are more broadly deployed if you count the number of unique public cloud service providers as your metric.

An API from a sales point of view isn't important because you're not selling an API. You're selling the applications that sit above the API and mostly those applications don't really care what's underneath. As a cloud service provider you're selling a value proposition, and unfortunately an API provides little inherent value other than potentially some reduction in development time if you decide to leave. Actually the really hard stuff is in moving Amazon machine images away from EC2 in a consistant way, which Amazon through their AMI format have made a practically impossible mission. [Paravirt, really?]

I'm not saying API's aren't important for cloud computing, just that with the emergence of meta cloud API such as LibCloud, Jclouds and others, programming against any one single unique cloud service provider API is no longer even a requirement. So my question to those who would have you believe the AWS API is important is again -- why? Is it because your only value is that in which there is little other than your API support? Or is there something I'm missing?

#DigitalNibbles Podcast Sponsored by Intel

If you would like to be a guest on the show, please get in touch.