The Australian Cloud: Is it delivering on its promises?


feature If you were to look back over the past ten years and try to pinpoint a specific point in time when the cloud computing paradigm specifically promised to revolutionise the future of computing, you would have to pick the two year period of 2006 and 2007. Up until that point, the term cloud computing wasn’t really being applied to a whole paradigm.

Before those years, isolated companies focused on cloud computing definitely existed — was founded in 1999 and NetSuite the year before in 1998 — and most of the major technology vendors were starting to get a feel for the idea that a different form of computing might influence their future development paths in a big way. But an industry-wide movement didn’t really exist yet. It would take several major events for that situation to change.

The first was the beta launch by retail giant Amazon in August 2006 of the company’s Elastic Compute Cloud (EC2) service. It was poorly understood at the time and panned by much of the media. But EC2 would go on to spawn a clutch of other cloud computing services over the next eight years and be adopted almost universally by tiny startups and major corporations alike. Amazon’s launch notes also contained a very precise definition of the promise of cloud computing that is useful for this article. At the time, the company said:

“Elastic Compute Cloud (Amazon EC2) is a web service that provides resizable compute capacity in the cloud. It is designed to make web-scale computing easier for developers. Amazon EC2’s simple web service interface allows you to obtain and configure capacity with minimal friction. It provides you with complete control of your computing resources and lets you run on Amazon’s proven computing environment. Amazon EC2 reduces the time required to obtain and boot new server instances to minutes, allowing you to quickly scale capacity, both up and down, as your computing requirements change. Amazon EC2 changes the economics of computing by allowing you to pay only for capacity that you actually use.”

You can see here all the key promises of cloud computing encapsulated in Amazon’s definition of EC2. The service would give customers quick and easy control over computing resources that ran on someone else’s infrastructure. These resources can be scaled up and down on demand; and customers will only pay for what they use.

The second major factor, and the one which sold the concept of “cloud computing” to business audiences, was the rapid growth of The company famously founded by Oracle executive Marc Benioff in 1999 had been steaming along throughout the early 2000’s, having survived the first dot com crash, but had not really hit its straps yet. And, although it had earned itself an enviable reputation for upending the market for CRM software, it had not yet really pushed too hard into extrapolating its CRM vision into a broader cloud computing revolution.

We have only been able to put together this feature series due to the support of its sponsor, Microsoft. With this in mind, please click to watch this brief video advertisement about Microsoft’s Cloud. The ad is only 31 seconds long and plays in your browser, but watching it will really help support Delimiter.

But by 2007, that had started to change. hit a key milestone in late 2007 of one million paying subscriptions (seats) using its core CRM platform. Benioff celebrated the milestone in a media release at the time and noted the tipping point had arrived. He said:

“Our strategy of delivering both the Platform-as-a-Service and Salesforce core Software-as-a-Service applications has driven a remarkable shift in the industry over the last eight years. It took seven years to reach the first 500,000 paying subscribers and only another 16 months for the second 500,000. That’s incredible global momentum and adoption for businesses of all sizes. For a million subscribers around the world, it’s truly The End of Software.”

If you take these two events together, what they add up to is a remarkable promise: The idea that cloud computing could end the need for traditional software by delivering everything businesses need through a flexible standardised service instead, with customers paying only for what they used.

But has cloud computing delivered on that promise thus far in the Australian market? Let’s take a look at the situation and find out.

In the previous feature article published in this series several weeks ago, Delimiter examined the topic of how Australian organisations were using cloud computing. In that article, we found that not all cloud computing technologies had been adopted with equal vigour. The most popular categories for adoption were Infrastructure as a Service (compute and storage resources) and certain categories of Software as a Service, especially CRM and collaboration.

In these areas, there is a substantial weight of evidence to show that, implemented correctly, cloud computing can deliver on the kind of promises which Bezos and Benioff were outlining in middle of last decade.

The Commonwealth Bank has emerged as a strong adopter of cloud computing services, with its own internal private cloud as well as using IaaS services from companies such as Amazon. And the results have been strong. In November 2012, Commonwealth Bank chief information officer Michael Harte told ZDNet that the bank had managed to halve costs across certain areas and was looking at savings of at least 40 percent from cloud computing services.

“We’ve halved storage costs, we’ve halved most of our app testing and app development costs,” said Harte. “As a general rule of thumb, we’re looking for 40 percent improvement in pricing across all the things that we consume as a service.” And it’s not just about cost savings. Harte also noted that it would have previously taken three weeks to provision a new server for a project. That time had been cut down to 90 minutes. In terms of its use of Amazon Web Services’ cloud computing infrastructure, the bank is also savings substantial amounts of money.

Local publishing company Ninemsn has also saved substantial amounts of money from shifting some of its processing load into the cloud, as well as gaining extra flexibility. “The publisher viewed the public cloud as an opportunity to bring the management of its systems back in-house,” iTNews reported in July last year.

There is a clear pathway in the IaaS space for organisations to achieve the full promise of cloud computing.

It generally starts with what might be termed ‘advanced datacentre modernisation’, with organisations virtualising as much of their internal datacentre infrastructure as possible. As that percentage creeps higher, additional functionality starts to be added to allow virtual instances to be self-provisioned, managed and even billed to different internal stakeholders using them. As this process of internal datacentre development progresses, organisations then often start to look at how they can shift some of this increasingly standardised infrastructure off to public clouds.

Usually organisations start with shifting testing and development workloads off into an external cloud. When this has proven successful, they will often start to shift some less-critical production workloads into external clouds — for example, public-facing websites. The Federal Government, for example, recently announced a plan which may see some 450 Government web sites shifted into a hosted platform using the Drupal content management system.

Some of the aims of Westpac’s own private cloud deployment detailed in late 2010 were the ability to quickly provision new processing capacity when needed, and cost savings gleaned through avoiding capital expenditure for each new internal technology project that required computing power.

Then-Westpac chief technology officer Sarv Girn said at the time that bank had already cut down the time it took to provision a new server “from weeks to days”. The organisation has first shifted much of its testing environments into the cloud, with Girn pointing out that it couldn’t afford to build a complete new traditional processing environment whenever a new technology project needed to be tested. “There’s a big application in testing, a big application in development, a lesser application in production,” said then-Westpac CIO Bob McKinnon at the time.

In May this year, a much smaller bank, ING Direct, revealed it had switched its entire production IT infrastructure onto a private cloud platform, in a move the company claimed was a first for any bank in Australia. The project, based largely on Microsoft software, with especial use of the vendor’s Hyper-V virtualisation infrastructure, represents the potential end case for major Australian organisations — with all infrastructure, even the most critical core components, completely virtualised and eventually shifted into some form of cloud platform.

Another aspect of the IaaS side of cloud computing is one that hasn’t necessarily been hyped and isn’t ‘sexy’, but is incredibly useful to organisations: Disaster recovery. By going down the virtualisation path and implementing some kind of IaaS cloud redundancy (even a quick and easy public cloud), organisations are able to ensure at least a basic level of disaster recovery functionality where they may not have had any previously. Federal agency Comcare went down precisely this path in 2013.

It’s interesting to note that, as with the rapid speed-up of the server provisioning process, using cloud computing to address this problem was likely the only way it could have been addressed at an appropriate cost. Writing on the Department of Finance’s blog about the Comcare implementation, whole of Federal Government chief technology officer John Sheridan noted that “with the agency’s current cost constraints, a traditional data centre solution would not be viable”.

What we’re seeing consistently here is that the IaaS business case does stack up over time, and cloud computing does deliver on its promises in this area: Certainly in terms of the private cloud, certainly in terms of shifting testing and development workloads to the public cloud, and at least partially for some production workloads.

Steve Hodgkinson, the director of Ovum’s Government practice in Australia and New Zealand — and a former deputy CIO of the Victorian Government himself — makes a distinction between cloud technologies, which is about assembling customised solutions for dedicated purposes within organisations, and cloud services, which sees organisations use external platforms for their work.

Every organisation in Australia had adopted some form of cloud technology, Hodgkinson said, but gradually most were also turning externally towards cloud services. The analyst views IaaS as a natural evolution of advanced data centre modernisation, but is more excited about SaaS, which he views as “genuinely a revolution”.

“I don’t think there would be a single software company in the world that was historically providig shrink-wrapped software that doesn’t have a plan to convert their offering to SaaS,” he added. “SaaS is rapidly becoming the default way to buy software. In the future it will be the exception to buy on-premise software.”

We see a similar solid ROI case when it comes to the uptake of Software as a Service in Australia in certain software verticals. The obvious example is A notable example in Australia of the technology being deployed is at major bank Bank of Queensland, which revealed in October 2012 that it was planning to deploy the platform to front-line staff.

The bank said at that stage that the system was low-cost, would simplify procedures and free up staff to focus on other activities. An updated provided by the bank in 2013 has shown that the outcomes were broadly reached. And critically, BoQ chief executive Stuart Grimshaw said, according to iTNews, that the SaaS model would allow the bank to access technology that was previously only available to larger competitors. the deployment also saved money. This deployment is fairly typical of deployments.

If we examine all of these deployments, what is apparent is two things.

Firstly, in most cases the executives concerned did cite cost savings as one aspect of the deployment, as well as the simplification of existing processes. However, as the project was completed, it appears to have been the delivery of extra functionality not possible without the cloud computing model that actually turned out to be the more important factor.

In short: Traditional models of technology deployment were unable to meet the developing needs of the organisations concerned; so they had to turn to cloud computing models to address their requirements. When they did, they usually also received ancillary benefits such as financial savings.

Right now, in Australia, these concepts are starting to be applied to other areas.

The Queensland Government, for example, recently abandoned a traditional on-premise model for its whole of government email needs, instead turning to an email as a service model delivered from the cloud. The NSW Government is doing the same. The previous legacy model just wasn’t working for either.

Virtualised desktop instances (or ‘Desktop as a Service’) are blossoming around Australia as organisations realise they can provide desktop PC services to staff more effectively through the cloud. Security as a service is also booming in Australia, and Microsoft and Google are seeing increasing uptake of their Office 365 and Google Apps productivity suites. Storage 2.0 platforms such as Dropbox and Box are making their way into existing environments where existing corporate storage solutions were proving unwieldy and inflexible.

And organisations right around Australia are watching closely as the SaaS ERP deployments which have proven successful in the private sector through small to medium businesses (especially through using Netsuite’s cloud platform) are making their way gradually into larger organisations. A cloud deployment of ERP (SAP) at the NSW Department of Trade and Investment is particularly being watched closely, as it may show the way ahead for resolving a series of failed ERP projects in the public sector.

Unlike in the areas of IaaS and SaaS CRM, the long-term return on investment and success of these use cases is hard to judge; it’s not quite possible at this stage to say whether these types of cloud computing will deliver on their hype just yet. However, there is no doubt that Australian organisations are right now attempting to leverage their success in early cloud areas into neighbouring fields.

Microsoft Australia Server and Tools Business Group Lead Toby Bowers has been in a good position to witness the cloud computing hype to reality cycle, due to his company’s position offering both on-premise and cloud computing technologies to customers, as well as the tools to move workloads between the different platforms.

According to Bowers, the cloud computing paradigm has definitely delivered on its promises in three areas — speed, scale and economics. You can see the speed aspect in the rapid nature of deploying cloud infrastructure, the scalability aspect in the way that both startups and major organisations have been able to ratchet up their use of it quickly, and the economic benefits in the way that it has saved costs or even enabled organisations to undertake work they would not have been able to otherwise.

And Bowers has definitely seen a more rapid uptake of cloud technologies in the Australian market than in other jurisdictions, leading to the major cloud companies such as Amazon and Microsoft itself to establish local datacentre facilities to serve local customers. He’s also tracked the shift to IaaS and SaaS over the past few years.

However, the executive is also looking forward to a new class of applications which will fill needs Australian organisations may not quite realise they have yet. Bowers expects, over the next few years, the shift to SaaS and IaaS platforms to taper off a little over the next few years as a new category of cloud adoption emerges — Platform as a Service.

What characterises the IaaS and SaaS migration paths that are still occurring is that the cloud has largely replicated existing sets of IT infrastructure and applications in a new environment. SaaS platforms may be more efficient, more flexible and easier to deploy than on-premises applications, but they still fundamentally sit within the same paradigm.

Bowers points to a new class of applications built only for the cloud as changing the game entirely. For example, he points to Australian company Coca-Cola Amatil, which is a substantial cloud adopter. The company is working with Microsoft to build new cloud apps to deliver functionality to ‘smart’ drink machines that can automatically size a customer up as they approach and ascertain what particular brand of refreshment they might be after. “85 percent of new commercial apps in Australia are going to be built specifically for the cloud,” the executive says.

Going back to the premise of this article — whether cloud computing has delivered on its hyped promises, it’s very clear that the answer is both yes and no. In some areas, notably IaaS and SaaS (especially CRM software and probably collaboration software), there’s a very clear business case — you’d be crazy not to explore engagements with the cloud in these areas.

However, in other areas, including in other SaaS verticals, cloud computing’s promise is still being explored, and the final answer is not yet apparent. Then, too, an emerging class of applications delivered specifically for the cloud may call into question, eventually, our definition of what this computing paradigm really is, ultimately.

For now, Australian organisations are, as always, best advised to closely examine what their peers are doing and don’t listen too closely to the vendor hype. Technology vendors are known for promising the world. But it’s only possible to ascertain the point at which hype becomes reality by looking at what’s happening in the real world. As always.


  1. Great article Renai.

    Thanks for this and it’s predecessor a few days back.

    Just found delimiter and think I am hooked!

  2. I found that example confusing – what is the difference between applications developed ‘for the cloud’ vs traditional applications that are deployed to and delivered by the cloud?

    • An example of a traditional application deployed to and delivered by the cloud might be installing SAP in a virtual machine in a private or public cloud infrastructure.

      The next step is Software as a Service, which means applications delivered purely on a cloud basis, through a web browser.

      But what Bowers is talking about is custom applications which are specifically developed to take advantage of Platform as a Service capabilities. So, for example, an application which may draw on a Database as a Service platform. Apps which integrate native cloud capabilities within them.

      Hope this helps!

  3. Personally I believe cloud technologies are inherently discrimatory currently – sure, SMBs can get access to infrastructure that provides platforms previously inaccessible because the cost of entry was just too high. But solutions like VDI and cloud storage for businesses needing to update a large (in gb) number of files daily are currently only usable for medium to large businesses that can afford the astronomical cost for dedicated fibre. Except of course for that tiny number lucky enough to be connected to the NBN. Until such time as we have Internet bandwidth available that is at least equivalent to internal network capabilities cloud solutions will be artificially held back, a situation that will adversely discriminate again SMBs when they would otherwise have stood to reap substantial benefits from this new level of scale and cost effectiveness.

    • I agree that you’re right — cloud applications are inherently bandwidth-limited. There are some (especially, as you mention, VDI) that are only achievable at the medium to high end enterprise level in Australia at the moment.

  4. @HonestlyCurious, herein lies a lot of the argument for the FTTP NBN model – exponentially higher upload speeds that allow massive volumes of data to be synchronized between local hardware and cloud servers quickly.

    While we wait for the speed of the connections to improve, software vendors have tried to “lighten the load” by developing systems that allow the data to “stay in” the cloud, rather than have to be downloaded/edited locally/re-uploaded. The rise of browser based systems (Google Docs, Office Web Apps, Xero, SalesForce, DynamicsCRM) that only require pulling small amounts of data from the cloud for the bits you edit on a web page, and terminal/remote desktop/streaming services that effectively just stream footage of the data instead of sending the files themselves all allow a relatively modest internet connection to perform a hell of a lot on the cloud without needing a big ass fibre connection.

    This method of cloud technology also allows for remote workers (working from home etc) where a business can basically reduce their own bandwidth requirements for cloud computing by having their workers use their own independent internet connections.

    Sure, there will be situations where you need to move a big ass block of data around the web, and unless you have a fat upload pipe, that’s gonna seriously suck… But the idea of cloud computing done right is that these sort of events become few and far between, so everything can just “work”.

    … At least until the Internet drops out :-P

    • “software vendors have tried to “lighten the load” by developing systems that allow the data to “stay in” the cloud, rather than have to be downloaded/edited locally/re-uploaded”

      Precisely! :)

  5. As someone that runs a small agency developing “SAAS” style apps for business, the “cloud” has been huge.

    We no longer have to worry about provisioning servers, patching, updating, etc. There’s a huge cost saving in this (in no longer requiring management for a lot of these), but the time savings are huge as well. No longer have to wait a week for a server to be provisioned, it’s one command away.

    Furthermore, if a member of our team wants to run a quick test, they can provision machines themselves to do it.

    We’re looking at taking this further, with offloading processing of our test suites to remote machines which are shared amongst the developers. We can then power these down after hours, so only paying when in use. Things are coming full circle again with thin clients and remote compute.

    Now hurry up and finish up that AU POP MS!

    • I hear this refrain constantly from small businesses (and I know a lot of them, being in a small business myself). Cloud computing has really unlocked a lot of capabilities for a lot of people.

  6. Cloud computing is a boon for big, medium-sized and small businesses alike. And there is no doubt that the Australian Cloud has been able to deliver on what it promised. Since all your data is stored and protected in the cloud, software vendors have successfully made it possible to improve the speed of Internet connections. We can always expect this technology to make things even easier for us in near future.

Comments are closed.