A desktop computer, or a bank of on-site servers can eventually carry out a large task. But many people aren’t prepared to wait the hours, weeks or even months that some jobs could take; while simultaneously tying up that computer at the same time.
The cloud, as in web, does in theory have oodles of surplus computing power. Power that theoretically should be available to anyone at the end of a high speed link up, but as with anything that is just there, reliability and guarantees don’t exist.
Enter the new breed of computer power providers such as Amazon or or Windows Azure, who are now productising extra processing power yet still are not solving the needs of your average user.
The Internet for most is about paying for what you want, when you want it. One of the problems with accessing this through these simple platform or server farm owners is they often want to charge in hourly chunks, even if a job itself only takes a few minutes. They generally haven’t designed it to be used on small intervals of time.
Naturally, this makes the would-be users of such Internet grunt more than a little bit gun shy about using the remote computing power.
“The people with the online computing power aren’t down to the granularity, or time intervals that most people need,” says GreenButton chief marketing officer, Vivian Morresey.
“Our software sits on top of top of these server farms’ platforms that enables micro-billing and true elastic or ‘on-demand’ provisioning. Our job is to better amortize, and aggregate the cost for someone wanting to use them to a simple rate.”
The Wellington-based company’s core piece of intellectual property is ‘Job Prediction’.
This enables GB to say how long a job, say a scene rendering, complex spreadsheet analysis or massive data interpretation exercise, will take and how much it will cost. Currently everyone else charges you for what you use, which sounds great, but how do you know how much you need?
“Job Prediction is essentially an online quote for your job,” says Morresey. Depending how quickly a client wants a task completed, they can receive what he describes as a ‘Goldilocks’ pricing. “If you want it in the next hour, it will cost this much, tomorrow, this much, next week, this amount,” he says.
Morresey says the company’s changing the whole nature of “true, on-demand computing” to the way that humans think about having an idea of what a job is going to cost – much the same way we like to know about how much a repair to our car is going to cost at a garage for example.
GB’s team of developers has created ‘middleware’, something that sits between a clients computer (say a design company), and connects to a range of servers and computers around the world.
The company has worked closely with Microsoft, which is well aware that its stranglehold in computing through desktop software is eroding as the world embraces the cloud. Microsoft has spent US$4 billion establishing many global server hubs around the world; accessible to anyone with high bandwidth.
Naturally Microsoft is keen for people to start using its server farms, and GB’s product is one way for it to participate in the cloud. Microsoft also likes it, because GB sits on the evolution path from desktop application to cloud application – allowing existing desktop applications to utilise cloud computing without a major rewrite of the code.
Under the GB model, at the figurative push of an on-screen button (push the green button, get it) a big job can be pushed out to the cloud, these individual parts of the job carried out across perhaps hundreds of computers spread around the world, and then reassembled back in the client’s computer.
The easiest way to explain it is to give an example of a ‘rendering’ job – where a simple computer model, say of a car driving down a street, is built up with its texture, lighting and shading all correct to produce a photo realistic moving image.
Using GB, that sometimes quite massive computing job, can be priced beforehand, and be delivered back to the home computer in the time required by the client.
Morresey says while GB’s software sits on the desktop, it takes advantage of the cloud.
“People will do more and more there, less on their desktop, and that’s one of the reasons Microsoft’s been interested in us,” he says.
The company has worked with Auckland-based IT company Right Hemisphere, and its ‘Deep Exploration’ application that lets users manipulate and ‘look’ at CAD drawings; zooming in and through multiple drawings which effectively turns line art into data entities.
As yet, nothing similar exists on the market, and the company sees its job prediction, and allocation and reassembly software as being applicable to the design and movie industries, financial analysis, industrial engineering and biotechnology/gene sequencing.
The analysis of oil and gas data is another use.
“People are now learning about other ways to get performance out of the internet,” Morresey says.
“We’re just at the beginning of thinking about all that. The timing is right for our product given the internet’s development and the availability of online computing resource. The mindset of people to trust the internet is also there, and they’re prepared to spend money over the net.”
Morresey says people are now more comfortable providing credit card details, or paying through means such as PayPal and that these are GB’s payment method.