Interview: Understanding the Big Blue cloud

When it comes to defining 'cloud computing', IBM is both revisionist and futurist.

As organisations like Amazon and Salesforce.com continue to take business away from traditional vendors, the approach of many a vendor product marketing manager has been to obfuscate the meaning of the word ‘cloud’ such that it no longer refers to a business model, but a box.

IBM – to the credit of its more conservative marketers – hasn’t been nearly as guilty as some of its industry peers. Being that 'utility computing' and 'on-demand' have been buzzwords in the company’s lexicon for some years, it’s perhaps understandable some at IBM would bristle at young web start-ups claiming to have invented the model, just as the entrepreneurs at the start-ups despair at phrases like 'private cloud' and 'cloud-in-a-box'.

This week, I spoke with IBM Australia's distinguished engineer Michael Shallcross about what IBM offers in this space, and what the vendor's future might hold.

Michael Shallcross, IBM Australia

Modern History

IBM currently attempts to play a role in any workload, via any delivery model, of enterprise computing.

First, there is the traditional vendor model, in which IBM sells customers the components (servers, storage, switches, software) to build their own infrastructure.

Then there is the traditional systems integrator model, under which IBM builds an infrastructure stack for the customer to own and operate.

It also operates under the outsourcing model, by not only building that same stack, but also maintaining and operating it on behalf of the customer.

And finally there is what might genuinely be defined as ‘cloud computing’ services - compute, storage or software offered as a service, over the network.

Like HP and countless others, IBM wants to label all of these services as 'cloud', being that they are all based on the same common reference architecture.

This ‘cloud-washing’ effort shields both vendors’ decades-old product, integration and outsourcing businesses from the cheaper (and sometimes nastier) public cloud compute services taking the industry by storm.

Today

Today, IBM's Australian customers can use virtual servers hosted in IBM data centres as part of the vendor’s six-year-old VSS (Virtual Server services) offerings.

VSS offers the provision of virtual servers (based on VMware vSphere 4.0) with a baseline configuration of one Virtual CPU, 2GB RAM and 30GB storage.

It isn’t pitched at the software start-up seeking rapid application deployment; rather, VSS targets large corporate and government customers seeking a slightly more flexible approach to managed IT services and outsourcing.

VSS falls short of the NIST definition [pdf] of a 'cloud service' in that customers can only engage in three to five year contracts.

It offers a two-to-five day window for adjusting virtual server configuration, a task completed by IBM staff rather than via the web browser.

IBM profits by offering services on top of VSS – such as monitoring and helpdesk – services Shallcross said has a “more manual involvement than pressing a button”. These extend to everything from management of operating systems, database (SQL, DB2, Oracle), Websphere app server, right up to full application management.

“A lot of our traditional outsourcing customers have already started using this infrastructure,” Shallcross said, highlighting particular interest from retail and manufacturing sectors.

Shallcross said the “less sexy” end of enterprise IT attracted most interest from IBM customers: long-standing managed security and storage services; remote backup of desktop and server environments; and shared data centre facilities for organisations to use as targets for disaster recovery.

IBM also provides a vendor-agnostic virtual desktop service – customers of which include some large public sector agencies.

Read on for IBM's plan for a cloudy future...

A SmartCloud future

Even with this depth of services available, IBM is acutely aware of the market share ‘cloud computing’ start-ups could potentially steal from its outsourced and managed IT services customers.

In March, IBM chief Sam Palmisano laid his bets on the table: cloud computing would be a US$7 billion ($6.5 billion) business for IBM by 2015.

Within a month, IBM launched SmartCloud enterprise – an IaaS service delivered from several global IBM data centres – the nearest to Australia being based in Singapore.

SmartCloud offers pay-as-you-go, self-service provision of virtual instances running Windows and Linux. It is pitched at application developers, marketers and casual users, with 99.5 percent uptime guaranteed by SLA.

The same platform will also be used to serve up middleware – tools such as Websphere app server, DB2 and Rational database tools, Lotus Domino server and some third party apps.

By the end of the year, Big Blue will launch a second generation "production quality" cloud labelled IBM Smart Cloud Enterprise +, extending its offer to the cloud provision of enterprise applications (such as SAP) running on enterprise operating systems (such as its AIX flavour of UNIX) on higher-grade hardware, matched with an SLA guarantee of 99.9 percent availability.

Shallcross told iTnews that this service would likely be based in Australia.

The bigger picture: Integration

But even with these two genuine cloud computing platforms released, the questions must be asked: Could IBM realistically hit Palmisano’s target of US$7 billion in cloud revenues by 2015?

Considering that the world’s largest cloud compute, Amazon Web Services, is tracking at around US$500 million in revenues today, IBM’s goal would be no mean feat.

But that’s assuming Palmisano’s US$7 billion is based only on revenues gained from SmartCloud.

IBM’s future role will be as an integrator and broker of cloud services. That future is already well underway after IBM’s acquisition of Cast Iron in mid-2010.

Cast Iron’s technology, Shallcross said, can "integrate workloads across both private and public clouds".

The US$7 billion figure also may assume that IBM will stretch the definition of cloud computing services well beyond what we know today as infrastructure, platform and software as a service.

Shallcross envisages that the SmartCloud platform might someday soon serve as the infrastructure stack from which IBM could launch a range of new cloud computing services.

What would some of these services look like, if we polished up the crystal ball? Shallcroft had two very interesting suggestions.

One was ‘business process as a service’ – under which customers might choose not just to have their application software hosted and delivered as a service, but also outsource the execution of the entire business process to IBM.

It wouldn’t be dissimilar to the way, for example, that Accenture manages the Navitaire airline booking systems used by the likes of Virgin Blue and Jetstar today. The difference would be that instead of a straight outsourcing agreement, or SaaS play, the customer wouldn’t pay per IT transaction but rather, per business transaction. It might be per reservation on a flight booking tool, per employee on-boarding on a HR tool, or per loan application on an online banking tool, for example.

Analytics is also an "enormously important play for IBM", Shallcross said, "a significant business opportunity" from which the company expects to generate a lot of revenue.

"We have started to look at SmartCloud as a platform to offer analytics services that multiple clients could consume," he hinted.

The catch is – competing with the Amazon’s and Salesforce’s would only reduce the premium IBM can charge its customers.

The challenge for Big Blue is to focus on areas where customers are prepared to pay for added value.