Sunday, August 17, 2008

IT as a Business: or IaaB

Whereas SaaS is predominantly a Business Process Outsourcing trend, Cloud Computing is predominantly an IT operation outsourcing trend. IBM and its shift towards services, supported by best-in-class hardware, has been leading this IT outsourcing concept for more than a decade now. However, IBM's main clientele have been large scale, high-availability, high-liability business such as Wall Street, credit card companies, and insurers. The shift created by SaaS as well as Cloud Computing is retargeting the IT supported business process services towards the mid-market where it will proliferate in a very different direction. High-availability is costly, as is high-security. Those are attributes that the mid-market is not going to pay for: they sure haven't paid for it in their own IT infrastructures. That will allow Cloud Computing service providers, such as Amazon Web Services, to provide a truly disruptive technology since it brings a whole new group of customers into the market. A segment that is cummulatively speaking more valuable than the high-end customers of yesteryear.

Jeff Barr, a Web Services Evangelist at Amazon, just published some interesting data that is supporting the observation that IT operation outsourcing is being leveraged aggressively. It also shows that this switch is happening incredibly quickly. For those in the business, this is not surprising because our customers have been screaming for more performance/capacity for a decade, mostly because the processor vendors such as Intel and IBM have not been able to provide anywhere near the performance improvements required to keep up with the data explosion.

From Amazon's 4th quarter earnings call, TechCrunch reports that the businesses that are taking advantage of IT operation outsourcing are not just tiny little start-ups:

"So who are using these services? A high-ranking Amazon executive told me there are 60,000 different customers across the various Amazon Web Services, and most of them are not the startups that are normally associated with on-demand computing. Rather the biggest customers in both number and amount of computing resources consumed are divisions of banks, pharmaceuticals companies and other large corporations who try AWS once for a temporary project, and then get hooked."

The value that is created by the on-demand capacity inherent to Cloud Computing is the big differentiator here for both startup and established business. For startups the value is inherent to the service, but for the mid-market it is on-demand capacity. To understand this, one must realize that most compute problems are bursty: it takes humans time to formulate experiments and setup the automation, but from that point on compute capacity and performance are the critical path.

In today's shift towards extracting more value out of operational business data, the mid-market is about to embark on a whole new degree of productivity; Cloud Computing and the pay-as-you-go business model removes IT operation, both CapEx and OpEx, as the limiter for all business: small, medium, and large. With this type of value creation, the switch-over can happen dramatically quickly.

Saturday, August 16, 2008

HPC is dead, long live HPC!

With the advent of web scale computing, genetics, and data mining for business intelligence the need for computer performance is growing exponentially. In a previous post, The Achilles' Heel of SaaS we identified that the processor vendors have been ruefully underdelivering performance so everybody has been forced to build clusters of cooperating servers to support the compute requirements of today's information processing. So now everybody is an HPC consumer.

If you look at the fortunes of pure HPC providers like Silicon Graphics and Cray it is obvious that corporate America has not been motivated by HPC vendor's marketing messages entailing the goodness of HPC for American's competitiveness. The only outfit that seems to keep these companies afloat is the NSA. This is a trend that has been documented for many years now at the Council on Competitiveness.

A quick Google trends query shows that just when everybody is becoming an HPC consumer, the term is slowly loosing its luster in favor of more business friendly terms like Cloud Computing and SaaS. SaaS in particular will force the provider to leverage HPC technologies such as clusters and distributed computing.

You can keep track of these terms here.

The data also shows that HPC interest and innovation has shifted away from the US to Europe and the far east. Organizations like India's Tata are building and operating world-class HPC installations. Even Sweden is on the top 5 list. It makes a lot of sense for Russia, China, and India to jump on this: they have very little inertia and they understand that moving up into the value chain is the next step of their evolution to play in the global economy.

Thursday, August 14, 2008

SaaS Business Process Outsourcing Buying Attitude

Citrix just published the results of a survey on the purchasing behavior, perceptions, and attitudes towards Software-as-a-Service. The survey results indicate that 15% of respondents are more likely to purchase a SaaS solution over a premised-based solution, if given the choice. That is 1 in 7, which to me shows that there is a lot of inertia in the IT space.

The first step in the survey was to gauge how familiar people are with SaaS. 31% indicated that they considered themselves knowledgable, whereas 50% felt less comfortable, and 19% had never heard of the term.

When asked which top five categories of applications they would consider purchasing as a service, Remote Access was number one (given that Citrix commissioned this study that is hardly surprising). The full rankings are shown in the following graph:

That only 15% of the respondents thought the SaaS model to be more attractive than on-premise software deployment is very telling. It shows that most people can't phantom switching processes/vendors/applications, and this is understandable. Very few software vendors make it easy to get data out of their systems, so creating the automation and validation procedures to transition is too much for many SMEs. Secondly, from a risk point of view, providing continuity during the switch-over is a daunting task, and something that most executives would rather avoid till the old system is so broken that it hurts. This is typically when the organization runs into financial trouble at which point reinvesting in something better is kinda the wrong time. Still, all understandable

Most SaaS providers understand this dilemma and provide lots of tools to serialize data from and to different systems. There are also plenty of third party solutions that make this process much more palatable. Cast Iron Systems is such a provider which allows on-premise and on-demand solutions for these ETL tasks.

Wednesday, August 13, 2008

SaaS aggregation and experimentation

In his guest column, How Not to End Up as an Anachronism, Greg Olson, the founder and CTO of Coghead describes the SaaS dynamics and particularly the new SaaS startup-up dynamics. It is a wonderful read, in particular, if you like analogies. Here is a quick excerpt of his column:

"We are now at point where implementors of SaaS capabilities are being disrupted by newer SaaS capabilities. Services that are built largely from other services are a reality, and offer many clear advantages. The types of services that could be used in the creation of new services span the spectrum, from base infrastructure services to complementary high-level application services that can be composed or mashed up. Example services include: compute and storage services; DB and message-based queuing services; identity management services; log analysis and analytic services; monitoring and health management services, payment processing services; e-commerce services like storefronts or catalogs; mapping services; advertisement services; in addition to the more well-known business application services like CRM and accounting.

The move to SaaS applications built on SaaS is a much more profound shift than the move from on-premise applications to SaaS applications. The software industry is beginning to display characteristics that mimic the supply chains and service layering that are commonplace in other industries like transportation, financial services, insurance, food processing, etc. A simple set of categories like applications, middleware and infrastructure no longer represents the reality of software products or vendors. Instead of a small number of very large, vertically integrated vendors, we are seeing an explosion of smaller, more focused software services and vendors. The reasons for this transition are simple: It takes less capital and other resources to create, integrate, assemble and distribute useful software capabilities."

In essence, with BIG companies like IBM, Google, and Amazon committing billions of dollars in infrastructure, SaaS start-ups and their financial backers can experiment very quickly with new business offerings and see what sticks. This is a dream for a venture capitalist: they don't need to be experts in understanding the business, they simply need to apply a portfolio approach to have their basis covered: this is typical MBA fare and thus very easily understood by today's VCs.

The reason these big companies are investing so heavily is because they believe that the future is SaaS and Cloud Computing. When infrastructure investments are taking place by multiple, large and international players, you know that the business dynamics are about to change.

Sunday, August 3, 2008

The Achilles' Heel of SaaS: the processor vendors Intel and IBM

Whereas the current SaaS offerings are dominated by the automation of standard workflows such as CRM, HR, and Accounting, the demand for information processing is quickly outstripping the capabilities of the processor offerings by Intel and IBM. The SaaS space will become much more competitive as compared to the old software industry that relied on on-premise installations. In the on-premise days, an IT staff had to become familiar with the installation, configuration, and maintenance of that software instance. The IT department had to amortize this investment in skill and in doing so became more and more intrenched with a particular vendor. With SaaS, that skill is no longer part of the equation and thus customers can more quickly move to a 'better' workflow. This implies that the SaaS providers need to differentiate through product value more so than before.

All business SaaS workflows manage important information assets that need to be leveraged to its fullest extent for an organization to be competitive. The computes to leverage this information have exploded due to maturity of data base management systems but also due to the sheer availability of competitive information available on the internet. Most of the current database research is about query languages and systems for querying non-stationary data and federated systems. Fully automated machine learning systems like Google's web index are only the tip of the iceberg. Compute requirements for optimization techniques in inventory management and logistics are much more severe.

This explosion of compute requirements creates a huge problem. Since SaaS relies on multi-tenancy to offer profit margin for its provider the end user actually gets less hardware to work with. Secondly, the demand for computes handily outstrips the offerings by Intel and IBM. At the end of 2007, Google computes were up at the rate of 10,000 CPU years per month and that rate has been tripling every year since 2004. All this implies that we are in for a bumpy ride with SaaS providers having to scale out their data centers rapidly due to customer growth and compute demands and Intel and IBM not doing their part to keep up with the compute needs. This will inevitably lead to an increased rate of innovation on the hardware side that the successful software organizations will leverage to differentiate. Virtualization of these hardware assets will be the name of the game to remain nimble and not get stuck when you bet on the wrong horse.

Friday, August 1, 2008

Google and Amazon as Benchmarkers

One of the most interesting aspects of cloud computing is the opportunity for benchmarking. Instead of paying market research firms Gartner or IDC thousands of dollars, industry wide benchmarking data would be readily integrated in the workflow offered by a cloud computing provider. SaaS, because of its single workflow focus, will be in the early lead to provide this added value to its customers. As a matter of fact, when I am evaluating SaaS providers and find out that they don't offer benchmarking, I discount them for two reasons.

  • The core functionality I buy from a SaaS provider is the best-known-practice for the workflow they offer. We all know that best-known-practices involve some performance based feedback mechanism. Thus if a SaaS provider doesn't have benchmarking, they don't have the best-known-practice.

  • Given that benchmarking is a core tenet of SaaS, if a SaaS provider does not offer benchmarking it is most likely because they don't have enough customers signed up to generate reasonable statistics and thus their longevity is in doubt.

  • The next innovation in benchmarking is when Google and Amazon jump in. Google has already started with Google Trends, but the real business value is held by online retailers like Amazon. Amazon as part of its own offering already demonstrates lots of benchmarking prowess in generating customer focused result pages that bring together your query and purchasing history and that of other shoppers that appear to be similar in context of your current query/purchase. For retailers this is so powerful. Amazon as a cloud computing provider has its platform perfectly positioned to attracked more specialized retailers that would not have the scale to do proper benchmarking. There is no doubt in my mind that Amazon will leverage this incredible information asset and make oodles of money.