I was telling someone recently that before I joined Microsoft I spent the late 1990s running a small training company. The number of employees varied, averaging out at a dozen or so. I delivered training, did the business management, helped the win over customers and I looked after the IT. It was like doing two or three jobs.
I’ve been quite reticent about our “Business Productivity Online Service“partly because it takes a long and closely argued post to cover why, from an IT professional’s point of view, getting rid of your servers isn’t abdicating. (This is not going to be that post). But as chance would have it I was looking at BPOS again with my old job in my thoughts. B-POS sounds like it should be something… ”points of sale”, but it is Exchange,Communications server and Sharepoint provided as Pay-monthly “Cloud services”
In the training company we ran all our own IT services, but there’s no way I’d host my own web-server today: the sense of using a hosting company was clear before I left for Microsoft. The launch of BPOS gave businesses a way to get hosted Mail (Exchange), Presence & IM (OCS) and Collaboration & Document management (Sharepoint) for $10 US per month – or in round numbers £80 annually – per user. Comparing that with the cost of server hardware and software and especially the time that in-house systems took up, if I were running that business today, my head would say get rid of the servers. You can mix in-house and in-cloud servers; users keep the same desktop software which is crucial: you don’t give up Outlook to move your mailboxes to the cloud.
It needs a change of attitude to give up the server. If my head argued costs and figures, my heart might have come back with benefits like “You are master of your own destiny with the servers in-house”. But are you ? Back then we couldn’t justify clustering our servers, so if hardware failed – work would stop until it was repaired. Paying for a service in a Microsoft datacentre means it runs on clustered hardware, which someone else maintains. Microsoft’s datacentre is a bigger target for attack, but the sheer scale of the operation allows investment in tiers of defence. Small businesses tend not to worry about these things until something goes wrong, and you can always tell yourself that the risk is OK if you’re getting a better service in-house. But the truth is you’re probably not getting better service. As a Microsoft employee I’m used to having access to my mail and calendar from anything that connect to the internet – laptop at home, or on the move, any PC with web access, or Sync’d to a phone. I doubt if I would have set that up for the training company but it’s part of BPOS – even to the extent of supporting iPhones and Blackberries. Getting rid of servers could not only save money but give users a better set of tools to use in their jobs – an easier thing to accept now that I don’t run servers for a business.
Now if you’ve come across the idea of the HypeCycle (see Wikipedia if not) – I agree with Gartner that cloud technologies somewhere near “peak of inflated expectations” – in other words people are talking up “the cloud” beyond it’s true capabilities, and if things follow a normal course there will be a “trough of disillusionment” before things find their true level. I don’t buy into the idea that in the future scarcely any business will bother with keeping their own server, any more than they would generate their own electricity. Nor do I buy into the polar opposite – that very few organisations, and none with any sense, will keep critical services in the cloud – that idea seems just as implausible to me. So the truth must lie in between: the method of delivering services to users won’t change from one foregone conclusion (the in-house server) to another foregone conclusion (the service in the cloud), like so many things it will be a question of businesses asking “does it make sense to do this in-house”, and I think IT professionals will want to avoid depending on that question being answered one way.
This post originally appeared on my technet blog.