Does anyone really care about hardware anymore? That's the question posed by Drew Robb in this provocative post on Server Watch. He laments the loss of enthusiasm that used to greet every new hardware upgrade, whether it was new chip-level developments from Intel and Motorola or new server, storage or workstation releases from the top manufacturers.
All the action, it seems, comes from the software side of the house. Virtualization, business intelligence, customer and database management all seem to generate more buzz than the steady stream of multicore releases or blade systems can even hope for.
On one level, I think Robb has it right: Hardware has crossed the line into the commodity realm, where technical advances take a back seat to price and volume. As such, there's little fanfare to accompany the launch of new platforms. And with the continued growth of open source software, the underlying hardware needs only to power up and make it go. The only time it gets any notice at all is when it fails.
Most of the major hardware vendors seem to realize this. The IBMs and HPs of the world have made no secret of the fact that software and services will be the growth engines of the future. Just yesterday, we brought the news that EMC has ventured into the server-management game with a pair of software stacks governing configuration and analysis.
Much of the hardware development today consists not of pushing the technology to new heights, but in tailoring existing platforms to suit new vertical markets. Cloud computing is a prime example, with many of the leading vendors keen on developing server andstorage platforms that cater to cloud services, hoping to steal the thunder of startups such as Rackable Systems and Verari Systems that have already made a killing in this area.
And yet, there's still something missing in all of this. If hardware is dead, does that automatically make software the new king? I used to think the answer was yes until I read this blog from tech publisher Tim O'Reilly. In it, he argues that the old software paradigms are giving way to a new order as well: the cloud. Yep, the very element being fed by new hardware is creating the platform that will bring an end to the static software applications that hold sway in the data center.
O'Reilly specifies three types of cloud: the utility computing model, the platform-as-a-service, and cloud-based end-user applications. His argument is that the value of traditional software is migrating toward these new layers in which applications are driven not just by software but by network-based databases fueled largely by user input.
So where does this leave us hardware aficionados? No better or worse off than we've been all along, actually. The entire enterprise universe depends on hardware -- storage to hold the data, servers to process it and network devices to get it from place to place. It may not have the same sex appeal it once did, but it will continue to provide the ecosystem for software, applications, services and whatever else comes down the pike.