Software as a Service (SaaS)

We had a previous article about this, with respect to eFuture (EFUT), There is a bit of a hype going on about the SaaS business model, which is somewhat understandable as recurring revenues sound much better than just a one-off payment, but SaaS also has it’s limits.

First, to clear something up which might otherwise remain murky, most of the ‘SaaS’ model that EFUT engages in is not really SaaS in the sense that the software runs on their (EFUT) servers (although the bfuture site seems to offer a modicum of this). The bulk of the EFUT model is related to after-sale service, maintenance, upgrades, and the like. We could call this model ‘SaaS lite’

‘SaaS pure’ involves running applications on servers that do not belong to the client. There are a number of advantages:

  • The client does not need a big IT department (by far the biggest slice of the cost is not related to equipment, but the personnel needed to solve problems, maintenance, software, upgrades, and especially connecting legacy applications)
  • For small companies especially, the opportunity to outsource all of these problems can be a real blessing, as it involves specialist knowledge and personnel, would usually involve capacity utilization problems (personnel on standby until problem occur), and it’s a hassle.

However, modern corporations are critically dependent on ICT applications, and outsourcing these also involves risk. These risks have become more prominent with the so called cloud computing model (which is basically the same as SaaS, it’s having your applications and data running on third party servers)

Here we have a contrarian view:

  • I promise this will be the last column where I moan and groan about the folly of so-called “cloud computing.” Personally, I think everyone should put their collective eggs in the cloud-computing basket, because that means I’ll win when the cloud goes down.
  • Let’s just summarize the events of the past few weeks. The Google cloud apps go lights-out for 90 minutes. The entire Internet is hacked and shut down in Georgia during the Georgia-Russia showdown. Various people using Google apps get their accounts killed because of bogus accusations that they are spammers.
  • In the meantime Microsoft blows up Microsoft Money and will not sell it as a shrink-wrap product anymore, instead putting the software package on the Microsoft cloud. It’s all part of the “software as a service” scheme.Let’s look at the basic reasons why cloud computing stinks.
  • 1) Performance issues. I don’t care if you have 30-megabit-per-second service—you’ll get flaky performance from most online apps, especially if they’re popular. Always remember that your online speed is only as good as the speed at which data is coming at you: The application server may be swamped, and the various nodes along the route could become clogged, too. Nothing is ever as fast as the machine sitting on top of (or beneath) your own desk.
  • 2) Software is software. People tend to forget that software is NOT a service; the whole cloud scheme is a scam to lock users into a single product and somehow extract more money from them. Aren’t these companies making enough profit by selling 35-cent DVDs for $500?
  • 3) Dependency problems. What happens to you and your business when suddenly your account is closed by a fickle vendor who decides that you are a spammer or your bill has been delinquent once too often? One thing’s for sure, there will be nobody to call—and if there is, they[[those people? better]] will be in Bangalore and useless. And of course, how many tech companies even stay in business that long? What happens if they fold?
  • 4) Cyber-terrorism. What happens if the net is attacked and your entire cloud world is gone for days and days? It just happened in the Republic of Georgia, and it can probably happen anywhere.
  • 5) The Internet is ill suited for this purpose. The Internet—a network of networks—was never designed as a client-server architecture for general applications. Yes, it can obviously be used for this, but it’s already swamped with too much data now. What will the cloud load look like if cloud computing becomes the future?
  • 6) Kills innovation. Ask yourself why the heck will we need six-core, high-performance chips if the cloud takes over everything?
  • 7) Reverses individual control. If you look at the history of computers—from big iron, government-financed projects to distributed computing to desktop/personal computing—the trend has been consistently toward individual empowerment. This has meant personal control within a self-contained environment. Cloud computing overtly reverses this trend. It is not self-contained and does not empower the individual, but instead makes the individual dependent and subservient. Thus, in the strictest sense of trend analysis, it is simply counter-revolutionary and a loser by any definition. From my perspective it’s just another gimmicky idea with a cool name that is doomed to failure until it is redefined completely.
  • Numerous people will tell you that you are already doing client-server “cloud” computing, because the Net is used for e-mail, and you go to Amazon for books or you use Google to search for Web sites. Google is a client-server “cloud” app, no?
  • These are specious arguments, since everything you do online involves another computer. And these uses cited by cloud-computing boosters are for information-retrieval purposes, tasks done online as client-server apps because there is no alternative whatsoever to these mechanisms. If you had a Google-size multi-petabyte cache of the entire Net on your local machine, I suspect you’d search it directly rather than go online to Google.
  • The key to understanding where cloud computing fits into the scheme of things is the question, What’s the alternative? If there is an alternative that runs locally on your own machine, it will always be better. And in most instances, there is a desktop alternative—and that’s what people will prefer and use. Cloud computing is for suckers.

Ok, the cloud computing Dvorak is talking about is basically for consumers, in business, the model is much more common. However, some of the same issues apply, especially the dependency and the killing of innovation factor.

If all companies run the same (or at least similar) applications on hosted services, IT would basically cease to be a distinctive competitive element, it would become a utility like electricity, basically. Of course, this was the thesis of Nicolas Carr in a famous (or perhaps we should argue infamous) article in the Harvard Business Review in 2003 (He subsequently wrote a book on the topic).

An additional point for business is that some applications are just way too complex, like enterprise resource planning (ERP) software.

ERP systems are tremendously complex software packages which uses a company network and a single integrated database to capture, integrate, and consolidate data and information about the most important company processes, like sales, (inboud and outboud) logistics, manufacturing, accounting, human resource management, and the like.

  • CIOs are still plagued by complex system integrations, high cost of ownership and ERP systems that are difficult to use. They are also yearning for innovation from their vendor partners. And yet, CIOs appear unwilling or unable to divorce their companies from their ERP systems and try something new, such as software-as-a-service (SaaS) or open-source models.
  • The degree to which companies are wedded to their ERP systems comes as no surprise to Amy Doherty, CIO of American Financial Realty Trust, a $426 million publicly traded real estate investment trust. Doherty says that over the past 15 years, companies have attached more and more critical data, including financials, to these systems. “The tighter your integration with ERP,” she says, “the more critical it becomes to your business.” []

ERP systems alow companies limited freedom as to how to organize their business processes (unless they are prepared to write applications themselves), and often forces a good deal of change in these, although most packages (like SAP’s) contain a good deal of customization options (by just flagging options on or off, basically).

However, the technology research bureau Gartner argues ERP as SaaS is not imminent:

  • A recent report from Gartner throws a big bucket of cold water on software-as-a-service ERP hype, especially for larger enterprises. Gartner analyst Denise Ganly writes in the “SaaS Impact on ERP” report that enterprises’ dire need for a suite of integrated ERP solutions is not something that SaaS vendors can reliably deliver right now.
  • “Because of the complexity of ERP suites, SaaS offerings for administrative and operational functions typically have provided functionality that is confined to one domain, such as sales force automation, or one business process, such as payroll,” Ganly writes. “Thus, ERP SaaS suite offerings are still immature.” (To read about an on-demand ERP provider’s nascent efforts, see “PeopleSoft Vets Born Again: Can Two Legacy ERP Guys Get IT Executives to Buy into On-Demand Applications?”)
  • Some of her other findings include: SaaS ERP suites won’t be viable options for large enterprises during the next five years. “Except for use in two-tier ERP deployments,” she notes, “large organizations should ignore this space.”

eFuture does have it’s own ERP offering, but this is by no means as complex as those of the big boys, like SAP, JD Edwards, and Oracle. But offering ERP as SaaS, it’s probably a little early for that still.

One thought on “Software as a Service (SaaS)”

Comments are closed.