It would appear that I have unwittingly caused a furor with my recent statement that “For those who ignore the cloud and carry on regardless, in around two to four year’s time their solutions will NOT survive”.
This was quickly picked up by noted blogger Adrian Bridgwater at Computer Weekly who blogged “There’s a lot of cloud propaganda out there right now and, as Akka also points out, not all companies will be shifting their mission-critical business applications and databases off their own servers and into a vendor-hosted cloud offering tomorrow. The option to consider hybrid clouds that provide the flexibility for organisations to move at their own pace — deploying applications in the cloud as and when they feel comfortable — should perhaps be discussed more openly”.
Having opened up the discussion Adrian quickly had quite a few comments but one in particular caught my eye.
MadaboutDana stated “Oh please! The guy’s got an agenda, he’s bound to say something like this. The cloud’s amazingly useful, yes, but comparing it with DOS is disingenuous. Just take a good look at the issues people are currently having with this so-called Cloud…
First, the latter actually consists of a wide variety of disparate services offered by many different suppliers with their own server farms – so there’s nothing uniform, predictable or controlled (i.e. single-supplier) about it. Unlike DOS!
Second, services have a nasty way of changing or vanishing. Look at Microsoft’s blogging customers, suddenly and unexpectedly transferred to WordPress. Look at BT Workspace, suddenly cancelled and taken out of service (it was, it has to be said, totally useless). Look at Google Apps and the continued failure to provide some kind of unified access system (just try giving domain users access to your folders by giving them share permissions [i.e. not sending them e-mails but giving them authority to access specific folders, like any normal file system]), and you’ll soon discover the limitations of Google Apps.
All these services are completely different, managed in completely different ways and run by completely different people. Recent furore over not just Facebook’s but also Google’s privacy policies (and, more insidiously, copyright and IP ownership policies) indicate just how risky total reliance on this wonderful Cloud actually is.
As an SME we use the Cloud for all sorts of stuff, including collaboration and off-site storage. But rely on it? No way! Everything is ultimately hosted on our own servers, so if one of our Cloud hosts decides to change stuff, delete stuff, or go bankrupt, it doesn’t affect our business. That’s simple common-sense.
Until the Cloud is a lot more unified (shared standards, failover procedures, backup agreements, unified security systems etc.) than it currently is, there’s no way I would compare it to DOS or any other operating system, simply because it’s not even remotely evolved as far as that yet. Web 2.0 is still in the process of happening, it hasn’t actually happened yet – and as for Web 3.0, which some pundits are telling us is already here: don’t make me laugh!”
Never one to shy away from a discussion I quickly replied “As I was the one who started this whole conversation with my quote above I feel that i need to write a comment here. My message may have gotten a little diluted and I would like to explain what I meant by the comparison to DOS.
Firstly, the context of my presentation was about business applications and the cloud. I was presenting arguments that cloud adoption by enterprise customers brings new architectural challenges to business solutions and software development in general. Issues such as latency, security, multi tenant, transaction and locking, data compression, end-user collaboration, billing, collection, scalability and more require a new n-tier architecture that is designed to work with all these challenges.
I concluded that traditional client server architecture will not survive in the cloud era, and that’s where I compared it to the architectural change between DOS and Windows.
Looking back at the beginning of Windows adoption by businesses, many software developers dismissed Windows, saying it is just a nice user interface, very graphic while saying that their existing solutions had adequate UI and most importantly very good functionality. Later on people started to understand the significant of Windows so DOS application used the emulator to run, still resisting change. The bottom line was that few years later, DOS applications did not survive the Windows revolution.
If Enterprise development teams did not change the underlying architecture of their offering to the business, the business users in many cases rejected the solutions, or resisted it to the point of killing it. And it was even worst for ISVs selling applications to the Enterprise market, without a windows architecture it became almost impossible. This was the basis of what my comparison was all about.
Now, I can’t resists not answering about “my agenda”. Firstly I do have an agenda, I am representing a company which market and sell an application platform that enables developers to develop mission critical solutions. Our application platform enables both on-premise and cloud deployment (and also mobile deployment on multiple smartphone platforms). So for us, we have the same interest to promote on-premise solutions as we promote Cloud based ones (or mobile for that argument).
My comments were made after careful analysis of many meetings I have had with Enterprise CIOs in the market, and many ISV CTOs; replying to many RFPs; attending many sales meetings and in general observing the market. It is very clear to me that applications which would not support cloud architecture and deal with all the issues that come with that, simply won’t survive, very much like DOS applications did not survive the Windows paradigm”.
And so it stands I would like to take the opportunity to thank Adrian For opening up the dialogue and MadaboutDana for allowing me to clarify my point. Should I start anymore storms online in the coming week, don’t worry I will keep you all in the loop.