The GoPC.net Blog

Cloud computing is all about efficient IT infrastructure utilisation.

In traditional IT environments, we see a lot of isolated systems that are fairly similiar – ie a shared file (or data) server, a network, and some desktop clients. If you stand back from this, its actually inefficient. Those machines are often left running 24/7 – it doesn’t take a rocket scientist to figure out that there is a lot of idle computers (1E and Gartner hinted at a global cost in the billions in this report).

Another flaw in traditional IT environments is the lag time inherent in not having resources immediately available – a business has to balance maintaining a “buffer” of IT resources (staff, time and hardware) against the cost of those resources. Bernard Golden (Cloud computing will cause three IT revolutions [CIO.net]) talks about the frustration that users can feel when trying to get resources allocated (in addition to the culture shock that cloud computing is causing in the IT industry!).

A larger pool of resources is, in general, going to be easier to manage then a smaller pool. This is a basic resource management principle – matching resources to demands. The more efficiently that you can utilise those resources, the better off you will be (or so the theory goes!).

This leads to a natural “next step” of moving to shared physical resources. This is where cloud computing comes in – directly addressing the pain of matching IT resources availability to demand.

Bessemer Venture Partners, Byron Deeter and Philippe Botteri, co- authored an excellent white paper which was published this week on Sandhill.com (the IT investment capital website from the heart of Silicon Valley, Sandhill Road). A 1-page summary is viewable at PRNewsWire.com.

Having pioneered most of our own development and direction through trial and error over the last 15 years I can attest to having made many of the mistakes outlined in the article and learned those lessons the hard way. It’s a good reference paper for anyone managing, investing, or working in a cloud computing Startup.

As a counter point to this article, we don’t separate the cloud into layers of SaaS, PaaS, and IaaS. The GoPC.net cloud fully integrates these functions together as one system and in doing so we have been able to create one of the most open, collaborative, cloud platforms possible. Cloud is far more than just “web apps”.

During the lead up to the Dot.Com boom all the analysts viewed the ASP industry (**) as being series of Ven diagrams that showed no overlap between the different industry elements. You “had” to be a network provider, a software vendor, hosting vendor, but you could not be everything. Then after the crash the only companies which were left standing were those of us who were at the intersection of everything, proving to me that the strength all along was actually in our model – vertically integrating all levels rather than focusing on one narrow layer.

Today, GoPC.net visualizes the infrastructure layer to provide an environment similar to Rackspace.com or Amazon.com, it provides a platform layer where customers use a virtual desktop infrastructure (VDI) equivalent to say a Microsoft PC and the network servers (without Microsoft license costs) and then on top of this we deliver the applications and collaboration capabilities which SaaS enables existing desktop apps as well as web apps. So what I see is a trend where most software vendors now refer to themselves a “cloud solutions” but focus only a thin vertical niche (e.g. Salesforce.com focuses mostly on CRM), others like Zuora.com with their payment system focus on one thin horizontal niche. But I describe GoPC.net’s version of cloud computing as a horizontal platform which services multitude of vertical niches.

GoPC.net is an open collaborative cloud that can run almost anything (web app, desktop app, even Microsoft based apps if we have to), and we’ve added to this by packaging into specific products the target markets from individual consumers right through to the enterprise. We offer an open platform so depending the ticket an organization buys from us, they can run almost any application or technology as a cloud solution.

Graeme Speak
CEO / Founder GoPC.net

** Cloud computing is the new name for Application Service Provider or ASP, which it was called during the Dot.Com bubble from 1999-2001.

Finally, after months and months of testing and data migration, we are ready to cut over all users to the new GoPC supercomputer infrastructure.

The change over will occur progressively over the next 24 hours.  There will be a couple of brief periods where all users will need to be logged off. This will be the first complete outage of the system affecting all users since we migrated the data centre in June 2006. It’s a record we’ve become very proud of but we’ve got to move forward. There are so many improvements on the new system, not the least of which is the time to create a new user account which drops from the current 4 minutes down to about 10-30 seconds.  Login times should also halve.  The next stages which will follow from this upgrade will be the rollout of new supercomputing nodes into North America and Europe and this will see a huge improvement in screen responsiveness for users closer to those regions (particularly Europe and Africa).

We have done our utmost to minimise the impact of this outage.  There are some areas which we have not been able to test until the live cut over is complete and there is the potential that we will encounter teething problems in some areas.  We will rectify any issues as soon as they arrise.  If you do experience any teething issues after the next 24 hours please contact us through the contact from on the website or email (support [at] gopc.net).

Once the new infrastructure is running we will start announcing a lot of rolling improvements to the system.  There will be a lot more choice available to customers, and a lot more freedom within some of the account types which we will have available.  Many thanks to those customers who have been answering our surveys and offering constructive feedback.

Graeme Speak
CEO / Founder

The problem with being a perfectionist is that one doesn’t want to mention anything until the pie is completely cooked, taste tested, and absolutely ready for presentation. The problem with Internet technology is that it’s never finished . . . But after years of refusing to announce anything until its actually running, I’ve agreed to put out this bulletin of what we’ve got coming down the line, just before it actually goes live.

I created GoPC in 1995 and we are now many generations of GoPC technology further down the path. The complexity in each new release grows exponentially. What becomes simpler and more useful for the customer requires a massive effort, discipline and very deep structured thinking on the part of the developers. I want to say that the GoPC team is the most intelligent, committed and brilliant team of people that I’ve ever worked with. It’s these guys and gals who have created the most incredible advances in GoPC for this upcoming release.

A special acknowledgement is also due to the remarkable people who create and contribute to Open Source, the thousands of developers who are responsible for the applications we’re able to provide via GoPC’s supercomputing framework at no cost to our customers. It’s our goal to add value to the work of these peoples by making their products much more available to a much broader audience.

The upgrade we’re about to undertake is a Major rework of the whole supercomputer. It’s new hardware and systems (approximately 100 separate servers not counting the expandable front-end nodes which carry the bulk of our processing load), new network infrastructure, the latest application versions on the desktop and a host of new very cool technologies. We have always made extensive use of virtualisation, and there is far more fail-over capability than ever before.

One of the most exciting advances from the new system is our ability now to fragment the supercomputer into separate pieces and distribute these geographically. This upcoming version of GoPC will become a distributed supercomputer with a footprint on each continent, all managed centrally. Changes and upgrades are automatically reflected across the entire infrastructure – thus effectively creating one global footprint. For customers who currently experience high levels of latency to the GoPC supercomputer in Australia (particularly users in Europe, Africa and Middle East where Internet routes go via the USA and can often exceed 400ms) this will have a dramatic improvement on screen responsiveness. We are currently negotiating with interested parties for financing and partnership involvement to achieve this quickly. Technically it is now about 10x easier than it was last year.

The new GoPC will give more granular control allowing us to tailor solutions for specific customer groups. Our premium enterprise customers and perhaps others will have the power to tailor their own solutions. For example they may choose from a simple pick list which applications their organisation will have; select another layer of data encryption technology on disk, add and remove users daily, select more server side storage, or recover a selected file from a last months backup straight from a window on the desktop. Godrive (GoPC’s online Drive) will seamlessly integrate with local PC desktop and you’ll be amazed at how fast this operates.

We have had countless requests since first going live in Dec 2004 with our open source version of GoPC, for a command prompt, for shell access and for the ability for people to install their own applications. This will soon become available and we will also allow customers to operate a complete virtual Linux desktop instance in their own virtual data centre.

This major upgrade crystallizes about 16 months of Research and Development by our team. Because of the scale of it, it has taken several months to migrate customers data and prepare everything for the big bang cutover. GoPC has run at 99.999% uptime since going live in Dec 2004, the only complete outages we’ve had was when we physically moved the data centre. At each upgrade we’ve only ever taken down part of the system at a time. We’re anxious about the change over because there are likely to be teething issues in areas which we could not have foreseen. But we have to move forward and we will be doing everything we can to ensure that any issues that surface are resolved as speedily as possible. Once we’ve switched over to the new system we plan to wait a few weeks to ensure things have stabilise and then start rolling out these new solutions which we’ve been wanting to bring to the customers.

Oh, I almost forgot to say, that everything will be running a lot faster. New customer sign-ups which currently are taking up 4-5 minutes, should reduce to around 30-40 seconds. The new GoPC works like lightning. We hope you’ll like it.

The ETA is any day now.

Graeme Speak
CEO / Founder

Our latest release of GoPC is being made available to our clients now. It sits on new supercomputer infrastructure within our datacenter.

Our latest release of GoPC has been developed with a view to being able to rapidly roll out and scale GoPC to millions of users around the world. In order to do that, and to ensure we can respond to spikes in growth, we needed to develop our implementation of GoPC such that we can duplicate our Perth datacenter infrastructure incredibly quickly.

It also provides new capability for us to be able to customize our offering and provide greater flexibility to customers as well. This means we have more flexibility going forward which will mean we can customise solutions for particular markets.

Our customers will notice more speed. GoPC is already fast, but our new release makes it even faster.

In addition, new and updated applications have been added to the GoPC desktop. For those people who haven’t looked at GoPC before or not for some months, it is well worth trying GoPC either as a replacement to your current desktop software solution, or to compliment and provide a mobile solution to your existing desktop software.

GoPC is really starting to make a move worldwide and are in discussions with major partners to provide similar supercomputer infrastructure around the world including Africa, Asia, India, North America and Europe. So, this will be great for people who are after a fast cloud computing solution.

We hope you are as excited about our new release of GoPC as we are, stay tuned for more developments…

I’m guessing it’s actually a matter of ‘who cares?’ Although some purists might gnash their teeth to hear that!

But if I had to answer it, I’d say it’s not actually a question of what came first. It’s more a question of which is a subset of which.

In my opinion, grid computing is a kind or subset of cloud computing. Grid computing is scalable (that is, make it big or small according to your needs) computing power, usually used by businesses. And supplied via the web.

It’s a way of tapping into extra computing power in ‘times of need.’ Say you need to do a mass mailout of a large document once a month (think: a pdf newsletter). Just tap into the power of someone else’s grid for the day and off it goes in the blink of an eye – rather than tying up your servers for the next 12 hours. Of course you pay for it. An Amazon Elastic Cloud Compute offers this service (although they seem to be hijacking the name cloud computing), as does GoGrid (who are, btw, not affiliated with GoPC).

All very interesting, but how does this relate to cloud computing?

Some would have you believe that cloud computing is either a fancy name for, or a part of, grid computing. Others go into infinite detail to categorise various providers WITHIN the cloud computing realm. (Actually, we think cloud computing encompasses grid computing, but that’s beside the point.)

This is because cloud computing as a term is still up for grabs. Purists can argue until they’re blue in the face, but the fact is that cloud computing will encompass whatever the people say it encompasses. Whatever works best and hits the market first and touches the public heart in the most widespread manner.

Best to keep definitions broad, then. Cloud computing is computing power offered via the web – something you can tap into without knowing or caring how it works. You just want to know THAT it works. And therein lies the principle of cloud computing.

Services, platforms, power, apps, hosting, etc you can pick out of the cloud that is the internet. A cloud is fuzzy, misty, obscure and opaque – but that’s okay because it’s not about how or why it works. It only matters that it works and you can access it anywhere, anytime.

Cloud computing is a method of supplying technology on demand. Who cares where it comes from, or even what it’s called? So long as it’s there when you need it. That’s why services as basic as web-based email are also forms of cloud computing. It’s just that, as the services expand, we are now finding it necessary to have a name for this stuff. And just in time, because the expansion it’s undergoing is nothing short of phenomenal.

Graeme