I have written about Cloud before, as I believe it has a good premise, however as a Cloud enthusiast I have been intrigued by the way the market did not embrace the technology.
Cloud computing has become a major IT buzzword in the last ten years or so, with predictions that it would completely revolutionise the way organisations procure and consume IT services, but so far it simply hasn’t lived up to the hype. Organisations haven’t rushed to embrace the Cloud; instead, executives have argued that it simply isn’t secure enough for their case, and IT departments refuse to acknowledge that it’s happening to such an extent that it’s predicted that 35% of total IT spend could be out of IT’s hands by 2015. So what went wrong, and why am I convinced that it could be about to make a big comeback?
SaaS vs Cloud
Firstly, let’s make a quick reality check on something that can get very confusing: what’s the difference between Software as a Service (SaaS) and Cloud? In short, all SaaS is Cloud, but not all Cloud is SaaS. I intend to talk about the whole of Cloud: I like to integrate components, not isolate them.
A Cloud environment is one which delivers common applications, providing computing on a utility basis. It removes dependency on infrastructure, allowing you to focus on what you want to do rather than running hardware; it removes up front capital costs (CAPEX); and abstracts the hardware upgrade loop because it’s no longer your hardware, or your problem. The characteristics of Cloud computing are: elasticity and scalability; self-service provisioning; standardised interfaces; billing and service metering. Confusing? Not at all:
- Elasticity and scalability means that the resource at your disposal gets bigger and smaller depending on what you need; it’s like having a two-seat sports car when you’re commuting to work which expands to a people carrier when you need to take your family to the beach.
- Self-service provisioning means that it’s easy to get what you need. Rather than a lengthy procurement process, just put your card details in online (or just log in and select what you need, if your company already has a subscription) and you’re up and running. It’s the difference between filling your car at a petrol station, and having to book a tanker to come to your house and fill it.
- Standardised interfaces simply mean that the services should have common handles which tell them how to talk to other services. This makes it easier to link systems together to not only replicate existing workflows but streamline and automate them.
- Billing and service metering means that you only pay for what you use, and are billed accordingly, just like with your electricity and water meters at home.
So, what, you ask, is the difference between this and SaaS? SaaS refers specifically to software, whereas the characteristics of Cloud can be applied just as easily to infrastructure (IaaS), where you get servers and storage as you need; platform (PaaS), Salesforce.com is becoming a great example, where you can write your own apps that work just like the original, with all the functionality you need; or even business processes (BPaaS) can be abstracted and procured as a service.
Mistakes made in adopting cloud
So what has gone wrong: why is it that this new paradigm just hasn’t taken off yet? The single biggest problem is that organisations had an immature understanding of the Cloud: they tried to adopt it as an upgraded version of hosted services. What exactly, you may be thinking, is the problem with that?
If so, you’re in good company. Even many cloud based businesses failed to market the new capabilities of Cloud effectively, instead selling web-based services on a per-user, per-month seat-allocation basis. So the only difference between “Cloud” and your existing on-premise or hosted software became the way it was being run, and the advertised cost savings.
We’ll come back to the cost angle; the deployment method alone was enough to put a dampener on the Cloud, because customers started to take the services apart, trying to understand whether it was secure, wondering if a “private cloud” (which is what we used to variously call a virtualised data center, thin-client computing, or even a mainframe) might not be a better option.
Is it actually cheaper? Or, why do you want it in the first place?
Then you get to the concerns over whether Cloud is actually cheaper than traditional outsourcing or in-house work. After all, you can simply take a look at your user count, estimate the costs to provision a service to them, and compare that to the Cloud rate card (or per seat cost). In fact, in many businesses the per-seat cost was vital as the existing processes didn’t allow for anything other than a fixed cost per month.
There are two problems with this: firstly, it creates a “race to the bottom”, where having decided that the target market is people who want cheap and cheerful, providers focus on delivering their services for the cheapest price, usually at the cost of user experience, reliability or security. Secondly, it ignores the real reason that businesses were interested in cloud: they wanted to just pay for services as they were used, get them in an easy way, and just focus on delivering value.
Remember the cloud characteristics I mentioned earlier? If implemented correctly, they provide different values to traditional infrastructure; as befits a new paradigm:
- Flexibility and agility: A good Cloud implementation allows your business to quickly get access to the services it needs to compete. I spoke to a major UK bank last autumn, who described their ability to create new products as a series of six-month waits to procure hardware linked together with short periods of productivity; unsurprisingly, top of their CEO’s wish list was the ability to bring products to market more quickly.
- Security: Ironically, given that security is the biggest concern around the Cloud, it can actually increase the security of your company’s data. Your data might reside off-site, but if it’s all encrypted and all that is transmitted to devices is images and keystrokes, is it really less secure than data sitting on a server in your premises? This all comes down to how well your service provider does their job, and how much it would cost you to employ the right skill sets and technologies.
- Risk: I’m always surprised how little businesses actually understand risk. Cloud technologies are encouraging a different approach to the risk of system failure; where backups and replication mean that not only are your services fault-tolerant, but you’ll barely notice problems.
Why I think the time is right for cloud
There is one key shift in enterprise IT at the moment which I think will drive much wider Cloud adoption: that is mobile, and in particular BYOD mobile. The reason is very simple; when mobile becomes a significant part of an organisation’s IT, additional investment in infrastructure is needed to cope, and this infrastructure needs to be available from any location, on any device.
It’s important to remember that services don’t exist in a vacuum, and whatever you choose to implement has to exist as a part of your employees’ workflows. This was the core of my recent blog on enterprise social media, and it’s no different in Cloud: if you don’t provide it as an integrated experience, it’s hard to be productive and could increase silos.
The two key additional costs are security and integration: how do you ensure your data remains secure; and how do you make sure this new tool becomes part of the workflow? Desktop in the Cloud is a perfect example; although there are many solutions on the market, they do not offer the cost-effective, elastic solution businesses need and the native performance demanded by end users.
Break the hype
As a provider of Cloud solutions, I have the conversations I’ve described above on a daily basis with clients who have been struggling to provide the services their users need. The good news is that we now have both the need and the technology to answer these concerns: mobile will be the driver, and as both our data and business logic workflows move to the Cloud we will use tools like in-memory data grid computing to overcome latency and provide robust, native-feeling cloud services to any device, anywhere.