Windows Server 2008 is about to be released. WHA!!?? It's 2008 already?
I really need to look back and take a look at my notes to see when I first was briefed about longhorn by MS…my guess would be in the 97/98 time frame, which would make it 10 years in the making. I also looked at Google trends (which only goes to 2004); longhorn interest peaked in 2005.
In fact, it appeared to me at the time of the win2K3 release that Longhorn was becoming the garbage dump for a lot of stuff they couldn’t fit into win2k3, but were under pressure to deliver….. The real question is can MS AFFORD to go through another product cycle like that, for an uncertain income stream????? And, more importantly, can end users??? The disruption for retraining/new technology just gets bigger and bigger as the legacy installed base gets huge with time.
Take a look at a mature industry like autos….there was a period where the technology was uncertain and significant new car models based on underlying technology was the rage (Steam, gasoline, diesel, crank, 4, V6, Wankel engine, etc etc)….now that innovation tends to be on the edges, and annual product cycles don’t typically don’t mean major changes in technology (that last major one was the hybrid), and certainly doesn’t mean significant retraining or service changes. I would even make the case that significant car model introductions happen every ten years or so, even though in the first half of the 20th century innovations came in 1-5 year cycles (any gear heads out there that would carry this analysis further, or challenge it?).
Economists would make the case that the governor for innovation is the human mind and its capacity to absorb new stuff – typically measured in 4-5% economic productivity growth per year, but shown to take a long time to ramp up over time (look at historical productivity statistics associated with computers). All that says is that change comes quick, early, then levels off while it is being assimilated by that ultimate end assimilator – our brain. That has held true with manufacturing, autos, and now computers.
Which all goes to say that MS better figure out how to add market-acceptable add-ons to the server that don’t take so long to develop (and charge for it)…I’ll bet they are trying to figure that out right now. Sounds like a SOA discussion at the OS level….Or, alternatively, just stretch the upgrade cycle out, accept that it’s now too hard and not in their best interest, and focus on the applications on top of the OS as money makers and introduce new ones (like SharePoint??).
Now, add in that businesses have another tool in their tool belt – Virtualization – and you could see that the pressures (and options) for businesses to not touch legacy applications (or delay any significant upgrades) are increasing. Does anyone remember IBM 7074 emulation under a 360, then 370, then PC? Insurers who jumped into automation in the 50s kept that stuff alive well into the 80s, and in fact it was only Y2K that really made them drop the last of it....
So my bet is the 5 year cycle will extend to 8-10, but it is questionable just how much value the next turn of the crank will have. I bet MS will try hard to figure out a better way to make money with the OS and change the form factor for shorter delivery and payment cycles...or just take it for what it could be: an unending support-based cash flow (ala Computer Associates strategy).
Jack