You are currently browsing the tag archive for the ‘Domain Specific Clouds’ tag.
In part 1 of this series I discussed the base technologies (virtualisation, shared resources, automation and abstracted services) that are at the base of cloud computing. Part 2 discusses the new computing models (Public Cloud, Utility Pricing, Commodity Nodes and Service Specializations) that have emerged as a result of the base technologies.
This part tries to understand the business value that can be extracted from these new models. After all, without value that can be easily understood by the business, there is little point in deploying cloud computing technology.
Part 3 : Business Value
Fail Cheap and Fail Fast
As a combination of a number of factors, the ability to try out an idea that can ‘Fail Cheap and Fail Fast’ facilitates the creation of business cases where the IT component does not become a burden if the endeavour is unsuccessful. In the cloud, if a business does not succeed there are no expensive paid for servers sitting around idle and no hosting contracts that are paid for, but unused, like gym memberships. In the cloud, the initial financial commitment is lower and the monthly burn rate controllable. If it does not work, you simply cancel the agreement and stop paying.
The ability of cloud computing solutions to handle growth allows time, effort and money on things that are more important during the initial stages, rather than on hardware and licenses that are going to sit around doing nothing for a while. It is common for the purse-holders, when receiving a request for budget to ask “How does this help revenue this quarter?” and planned, prudent and reasonable infrastructure purchases simply do not generate revenue until sales pick up. So having a platform in place to be able to demand additional resources when necessary negates the need for up front purchases. An important observation however, is that this only makes sense if growth is expected. A website that is expected to be small or self constrained (such as a corporate time keeping application that has a finite number of users) may be better suited to a Plain Ol’ Web (POW) app and forgo the cloud computing engineering costs.
The reason why Amazon is a cloud provider is because they needed a lot of hardware to handle sales during the Christmas season, which sat idle for the rest of the year, and this spare capacity started to be sold off on the cloud. Many businesses have similar situations where there is peak or cyclic (per day, season etc.) demand, such as the Christmas rush, or unpredictable demand, where the site is suddenly mentioned by Oprah. Peak demand periods are important for businesses. It is often the time when the first time customers, which have cost a lot of marketing money to attract, who visit the site and expect a positive experience. Cloud computing caters specifically and overtly to the handling of peak demand periods.
Due to features that are part of cloud computing solutions, a lot of risk can be taken care of out of the box, so in many respects cloud computing can be seen as part of the solution to managing risk – operational, reputational, disasters and so on. Although cloud computing security could increase risk many other fundamental requirements and features of cloud computing platforms, such as backups, availability, patching, load balancing, scalability and others (in an automated, zero touch manner) does tick some risk management boxes.
Time to Market
In a competitive market, a product’s development cycle and time to market is key to the viability and planning of a product. Having to pad the launch by a few months because of the provisioning of IT could scupper the entire product proposal. While IT has a tendency and history of not delivering on time, cloud computing can, in some cases, reduce the time to deliver, particularly if the alternative involves a long hardware, software and networking procurement process.
Because the cloud computing is about consumption of units of computing that are billed monthly (or daily, or some other period), the idea that computing costs are operational expenditure rather than capital expenditure is often touted as a benefit of cloud computing. While true and relevant in some cases, financial models and needs of businesses cannot be generally applied. Different businesses have different (and complex) financial models that may or may not find the capex of IT hardware a decisive issue.
Enterprise IT Backlog
Lurking within all businesses is the dissatisfaction in the rate of delivery of centralised IT which seldom has the skills and resource bandwidth to cope with the torrent of new business requirements and applications. Rather than having their particular needs sit for months or years in the enterprise IT backlog, disgruntled and impatient business units are taking their budgets to external organizations for fulfilment. The tradition of getting and external development company to develop bespoke solutions and force enterprise IT to install and support it will be replaced by development, support and operations being completely off site. Leaving internal IT in the dark and toothless. Salesforce.com has ridden this demand and many cloud providers will cater to these rogue bespoke solutions.
Domain Specific Clouds
Hollywood studios have a need to do a lot of CG rendering towards the end of movie production when time is running out. Having the necessary horsepower sitting around for when it is needed is expensive and quickly becomes redundant so studios hand rendering over to third parties that have huge capacities to take on particular jobs. While not cloud computing per se (for example, I am sure they ship data on rather large hard disks rather than use the internet) the idea of having specialised processing service that offer more than just computing power is beginning to be embraced into the cloud computing landscape and the term ‘Domain Specific Cloud’ is being tossed around. A more common example is data mining, which addresses a whole lot of services including forensics, fraud detection, deduplication and other value added services that are a lot more than just raw computing.
So what are these new ways of doing business that are emerging as a result of the value that can be realised from cloud computing?
Continue to part 4 : ‘Emerging Business Models’