It was beautiful when it was a concept, and it's even better now that it's a reality.
Let's step way back.
About 10 years ago, Larry Ellison appeared on the Oprah Winfrey show to discuss ''network computers.'' This was when a lot of ahead-of-their-time ideas were out there, like Apple's Newton and IBM's voice recognition applications. Larry was wrong then because our networks weren't reliable, secure or ubiquitous enough to support ''thin client'' architectures. Never mind that the devices themselves also were a little weird and way too proprietary.
But if Larry reprised his appearance on Oprah tomorrow, he'd be dead right. The concept was always right.
Let's look at several trends that point to why ''thinfrastructure'' makes sense.
First, we've segmented our users into classes based on their frequency and depth of use. Some people use very expensive and hard-to-support laptops pretty much for email. We've begun to deploy a bunch of devices -- like PDAs, cell phones and pagers -- that are converging. Many of these devices are pretty dumb, but that's OK because we don't ask much of them. Multiple device synchronization is, however, difficult and expensive. Desktops and laptops are way overpowered for the vast majority of users, as is the software that powers them.
Enterprise software licensing is getting nasty. For instance, if you don't renew your Microsoft software license every three years, you get punished. Desktop/laptop support is still complicated and expensive. Large enterprises, even if they have network and systems management applications, still struggle with updates, software distribution and version controls. And the larger the organization the greater the struggle.
Network access is ubiquitous today. We use desktops, laptops, personal digital assistants (PDAs), thin clients, and a host of multi-functional converged devices, such as integrated pagers, cell phones and PDAs, to access local area networks, wide area networks, virtual private networks, the Internet, hosted applications on these networks as well as applications that run locally on these devices.
These networks work. You can be on if you want to be on.
Many companies provide multiple access devices to their employees, and employees often ask their companies to make PDAs compatible with their networks. All of this is made even more complicated with the introduction of wireless networks, which make employees more independent and mobile, though less secure.
The cost to acquire, install and support all of these devices is out of control. What if I told you that the annual support cost for each wireless PDA in your company was more than $4,000? Youd either laugh or run and hide.
Run and hide.
How should companies approach the network access device problem? What are the viable alternatives -- today and three to five years from now? Should companies prepare to support multiple devices and access strategies for each employee? Should we standardize on one device? Should we skinny it all down?
We should think about thinfrastructure.
Small, cheap, reliable devices that rely on always-on networks make sense. Shifting computing power from desktops and laptops to professionally managed server farms makes sense. Moving storage from local drives to remote storage area networks makes sense.
Fat clients should lose some weight -- as we bulk up our already able servers. The total cost of ownership (TCO) -- not to mention the return on investment (ROI) -- of the skinny client/fat server architectures is compelling to put it mildly.
Is this about control? Is it about reliability? How about security? Yes.
Think about all of the new devices appearing on the market and how many of them are converging into wholes greater than the sum of their parts. Since we all agree that timing is everything, isn't this a great time to think thin? If you don't, then the only people you'll make really happy are the vendors you have to hire to keep all the toys humming.
If you get control of the servers, you control everything that plays on every access device. And if you skinny down the access devices, you get control, flexibility, standardization, reliability and scalability for your users. No more software conflicts, instant software upgrades, quick-and-easy new application deployment.
Sound too good to be true?