If the Cloud is just about cost savings, then it's pretty dull. It could be about so much more.
The reality behind the cloud is pretty boring. When you look closely at what most people are doing there, it's all about cost savings.
Take a commodity application from your on-premise servers and push it out to the cloud and your costs (probably) go down.
Few applications are heavily loaded 24 hours a day, seven days a week, so paying only for the compute capacity you need as you need it makes a lot of sense, but it's hardly exciting.
People have been renting resources rather than buying them for centuries. Cost savings are nice, but they're just the bread-and-butter of good business management. As the saying goes, no-one ever cut their way to greatness.
Occasionally people go a little deeper and start to look for "agility". In the cloud world, agility mostly means making it easier to bring up new servers as you need them. This has a couple of nice effects.
Firstly, it makes it easier to develop and release new applications – development teams are almost always constrained by lack of access to development and test environments. Cloud reduces this constraint.
Secondly, it makes some applications a lot more feasible. Large analytic applications that need a huge amount of hardware but only run intermittently are the classic example, they'd rarely be feasible if we had to buy enough hardware to service their peak demands.
Which brings us back to cost savings. This agility was always possible if you had the cash, cloud just brings it into the realm of organisations with budgets.
But is that all the cloud's about?
I don't think so. I think there's plenty of scope to push the benefits of cloud a lot further.
Some smart people I've been working with are now starting to look at how they can use the cloud to change their innovation lifecycle. They've realised that those initial enhancements to agility create an opportunity to completely rethink the way they develop new products and services.
Most organisations have some sort of product development lifecycle. This lifecycle may be explicit (documented in some dusty manual somewhere) or it may be implicit (hidden within the assumptions that drive "the way we do stuff here"), but it's nearly always a single flow.
New products and enhancements pass from ideation through design and development to release and customer support. In sophisticated organisations, people recognise that it's really a loop rather than a linear flow – customer feedback triggers fresh ideation and product development / enhancement.
Either way, the organisation's product development teams, project management processes, development environments, etc, are then all structured around this lifecycle.
For those smart people I'm working with, cloud has created scope to rethink that lifecycle. They've recognised that products go through several phases, each with their own lifecycle.
For example, they're starting to recognise phases such as:
- Pre-product:In this phase we're just exploring new ideas that might become a viable product. The pre-product is in perpetual beta with a small market of early adopters.
- Competitive product: Now we're taking the product out into the wider market, and delivering regular new features and versions in order to stay ahead of the competition.
- Commodity product: The feature set is now a lot more stable, and our emphasis has shifted to reducing the overall cost of ownership and operations.
Each phase has some sort of development lifecycle involving ideation, design, development and release, but they all operate within very different parameters. Pre-products have a cycle time of days and can tolerate relatively low (99%) availability.
Competitive products deliver new versions on perhaps a quarterly cadence, but require much higher (99.99%) availability. Commodity products launch new versions on an even slower cadence, and focus heavily on optimising resource utilisation and manageability in order to reduce costs.
On the face of it, that's hardly a radical insight (although it's way beyond most organisations' product development lifecycle). The interesting thing is that the cloud enabled, or at least triggered, busy people to start shifting their mindset so they could envision a much more flexible approach to product development than they were currently achieving.
They hadn't thought about this previously because the overheads of supporting multiple servers, development and test environments, etc, were too large to even contemplate.
Now the cloud is making it economically and managerially feasible for them to support pools of servers that operate within each of those different parameter sets. This in turn is shaking up dramatically the way they think about developing, market testing and releasing new products and services.
It's in shifts to mindset like this, triggered and enabled by the cloud, that the real excitement lies.
Graham Oakes helps people untangle complex technology, processes, relationships and governance. He is a guest blogger on Econsultancy.
No hay comentarios:
Publicar un comentario