For obvious reasons of their own, tablet makers (ahem, Apple) have long been trying to kill off the PC. Depending on how you read them, recent PC sales reports seem to suggest that PC sales are declining while tablets and smartphones are replacing them, giving credence to the tablet makers' claims. You can argue both ways. In fact, in The PC is Not Dead. It’s Just Different, I threw out my thoughts on the matter. Paul Thurrott has been covering the PC sales reports for a while and keeping pretty tight watch over what's transpiring. You can read through the long list of posts here: PC Sales Coverage. The only true bright spot in the PC manufacturing world is Lenovo. Despite the rest of the PC industry struggling, Lenovo keeps defying analysts. At TechEd 2013 Lenovo's ThinkPad Helix won the Best of TechEd for pure innovation and they've recently unveiled a new manufacturing facility in the United States. Lenovo seems to be the exception that all other manufacturers need to model after. It doesn't take too much straight thinking to realize that the PC isn't dead at all.
However, it has struck me curious recently there's been very little coverage on the opposite end of the computing spectrum: the server.
Today's servers are required to spin up a multitude of VMs quickly and efficiently. They are built to service thousands of end-user requests and have been streamlined for efficiency. So, today, it takes a lot less server hardware to perform the same functions that were required of it only 5 years ago. A single server today performs the same work cycles as 10-20 servers in a server farm did 10 years past. So, obviously, there are fewer servers being purchased. On top of that, servers are one of the last pieces of hardware to be replaced in a company. If the server is working, no one wants to touch it. IT is even reluctant to apply security updates in the event they will cause a downed server.
Now, we are constantly bombarded about the Cloud, how it's the future, and why it's important to migrate on-premise software and processes. Yet, as we replace on-premise processes with Cloud work cycles, what happens to the server hardware? It's no longer needed and if the servers are no longer in use, there's nothing to replace.
Digging into some research it appears that server sales have been declining for quite a while. In June of 2012, IDG reported that server sales had declined by 12 percent for Europe, the Middle East and Africa, citing difficult market conditions. As recently as May 2013, HP issued statements that Itanium server sales were down 37 percent, while Dell's combined PC and Server profits fell 79 percent. It's not clear how much of Dell's reported profits were PC and how much were server related.
Obviously, server sales are down and steadily shrinking, and, there's a two part message here. First, a single server, today, can perform the same job as a multitude could in the past. Secondly, as more companies move on-premise infrastructure to the Cloud, the Cloud will eventually be the largest consumer of servers, further pushing server hardware sales into the toilet.
So, unless Lenovo finds some way of innovating server hardware, the server market may tank long before the PC market does.
In large organizations there are entire teams dedicated to the upkeep, monitoring and management of servers. For the longest time I've heard the complaints from IT in companies that utilize System Center Configuration Manager for patching, where the "server team" won't allow the "client team" (the ConfigMgr admins) to touch the servers with any updates. Updates are tested, approved, and staged solely by the server team and if there's any deviation from the process, there's trouble. So, if servers become non-existent within the organization, what happens to the server team?
I'm interested in your thoughts. Is your company budgeting to add server hardware over the next year? Are you replacing server hardware costs with Cloud subscriptions?