Getting started with SQL Server Data Services (SSDS)

I spent this weekend primarily doing two things:

1) hacking away at my new HTC Kaiser (AT&T Tilt, TyTn II), unlocking it, flashing new ROMs, installing new CABs, and
2) getting familiar with the basics of using SSDS.

The David Robinson MSDN Magazine article that helped kickstart me on this topic can be found here. Even better, it turns out there's an invaluable SDK/toolkit for getting to know SSDS -- the SQL Server Data Services SDK (beta) of course. There is also a single MSDN forum dedicated to SSDS. I haven't yet discovered an IRC channel. There are a number of MSDN team blogs, as well as a handful of non-MS blogs that I'll cover in later posts.

To up the ante a bit, I decided to write a console (command line) app for my Windows Mobile 6.1 device, the HTC Kaiser. I was halfway through implementing this app when I discovered the SDK, and the command line tool that comes with it. I'm going to reimplement this tool, ST.exe, for Windows Mobile/.NET CompactFramework 3.5. (I've tried to find source for it, so far, no go ...)

The SDK also comes with a GUI explorer that makes it simple to understand the RESTful operations supported by Authorities, Containers and/or Entities. (You can't PUT, or update, Authorities or Containers. Yet attempts to do so still seem to increment those objects' version numbers ...)

So far, I'm liking the SSDS beta experience. More to follow.

Back-and-forth: Microsoft's cloud data offering, SSDS

Quick, raise your hand if you're writing code using SQL Server Data Services right now.

How about, raise your hand if you've even heard of SSDS?

Microsoft's nearly-spanking-brand-new cloud data offering, SQL Server Data Services (SSDS) is a Google BigTable-like horizontally-structured data cloud, still in beta.

Horiztonally what? The basic data model is authorities, containers and entities. Using authorities, one is able to geographically distribute one's data -- not unlike Amazon's Availability Zones, and something Google App Engine and Google's BigTable betas don't yet offer.

Truly scalable web applications beg for this sort of backend. For a slight trade-off in latency, you have extremely affordable, dynamic, distributed capacity.

My friend Michael O'Neill over at crisatunity, one of those DBA guys, isn't a big fan of the data cloud model, at least not at the oustset of this conversation. We're going to have a blog back-and-forth on this topic over the next few weeks, as we both test out the beta, and I prepare for a couple of presentations for an upcoming Code Camp and local user group.

Not only do I come at this from a Developer (vs DBA) perspective, but this sort of service is the sort of thing that's helping drive down the cost of starting up a new business. As CTO of Change Round-Up, I believe cloud services offer us the ability to address a Top 50 retailer's peak transaction volumes, without having a lot of over-priced under-utilized in-house capacity.

I look forward to seeing where this conversation takes us, and if anyone's opinion on SSDS, or cloud services in general, is swayed.