The other day Paul Neilson wrote a simple post making predictions about SQL Server in the Cloud. That got a response from Denis Gobo talking about a few things that need to happen before we can really take advantage of SQL Server in the cloud.
Most of my work takes place in the cloud. Actually it’s half my work, but I had an interesting experience yesterday.
I was working along when bandwidth slowed to a crawl. Pages wouldn’t load, even Google took a minute to bring up their lightweight home page. My tweets wouldn’t post, Outlook would freeze while checking email, etc. Normally that isn’t the end of the world since I do write quite a few things offline, using Live Mesh to synch files among my two different machines.
My process is typically to write offline, load the content into the SQLServerCentral site, and then once it’s up there, delete it from my local files so I can keep them pruned. Writing 5 or 6 editorials a week, every week, means that if I keep too much stuff around locally it’s very hard to remember what’s gone out the door. Maybe that’s not the best practice, and I’m actually rethinking that now.
In this case, however, I needed to get the podcasts prepared for the next day. So usually I open the data online and use it to shoot the podcasts.However I couldn’t get to my application online, so I couldn’t get the podcasts done! As the day wore on, and this was over hours, I stressed more and more, and ended up getting worried. Luckily I had one piece of content still on my machine, as well as a few more that weren’t scheduled.
So I worked with those, got work done, and moved on. Today bandwidth is better, and things are moving along smoothly. But I’ve seen bandwidth issues at companies before, and if you are too dependent, or don’t have backups, than what do you do?
If SalesForce.com goes down, it makes major news because it effectively slows or stops work from being done for a lot of customers. Is that any different than a server going down inside your company? What if your databases were hosted somewhere and things went down?
The advantage of the cloud is that there should be a bit more expertise and effort put into making things more reliable. There are many companies, including two of mine (JumpstartTV and End to End Training) that host their servers in a data center, relying on the Internet for access.
While I’m nervous, and I think the insurance industry really needs to become more involved in ensuring security is well developed here, I do see SQL Server moving to the clouds. I think there need to be a lot of companies offering this, like we have lots offering web hosting now. That has become almost a commodity offering now where you can easily move your application to another company if you have issues.
I’m not as aggressive as Paul since I think there are a lot of tools that need to be built and some maturity in the SQL Server platform, I think that a very full featured SQL Server offering is still 3-4 years away and I can’t see a large percentage of companies avoiding buying servers for a decade. I’d predict that we’ll see more “internal clouds” in companies in the next 3-5 years where you won’t necessarily connect to a particular server, but a URL in a cloud somewhere.