As I read this blog post on an interesting idea around how to get help up-skilling on the the cloud providers (AWS etc) i reflected on the rate of change we are seeing in the software market and what that means to administrators that are responsible for managing this change.
In the Web 2.0 areas we deal with we see change on a regular basis, some is transparent to me and some requires a little effort. A couple of examples:
- Xero does new releases on a regular basis, apart from a small outage window when Xero isn’t available I do nothing but login and start using it again.
- Our blog is powered by wordpress, its self hosted, so when I login as the admin to write a post it tells me there is a new version and a wizard upgrades it for me. A little more effort.
- Our website used to be powered by Joomla (again self hosted) and each major Joomla version (1.0 > 2.5 > 3.1) required us to implement the new version and migrate our content across, lots of effort.
When we look at the enterprise BI vendors software we have pretty much been stuck with option 3.
And when we are talking major underlying technology changes or major product component swap-outs, then things get even harder. Some examples:
- The move from Business Objects ver 6 to XI (found a table with all the version numbers here)
- Going from Oracle Discoverer to OBIEE
- Going from SAS 9.1 to SAS 9.2/9.3 (and now from SAS eBI to SAS Visual Analytics)
Luckily these mega-vendors are dinosaur slow and so new versions only arrived every 3-5 years. Which wasn’t so bad because often a customer would want to reimplemented everything from scratch using a new vendor after 5 years anyway.
But no more!
We are seeing vendors releasing maintenance or service packs regularly that add new functionality and with products like SAS Visual Analytics we are seeing new releases with major new functional changes every six months. Each of these requiring major effort to implement.
And when we look at providers such as Amazon Web Services (AWS), we a rapid cadence of change. Now the differences is we don’t have to do anything to get these changes, but it does mean we have to spend time evaluating them, deciding if they are directly impacting what we are doing today and if we should make changes to what we do tomorrow to do things better.
A great example is AWS Kinesis which has just been released.
- What is it?
- How does it work?
- Can we use it to replace our current data feed approach?
- Can we use it to do ELT more efficiently?
- Can we use it to save our customers money?
All these things an administrator needs to do, so the cloud is moving their role from not just being a person who maintains the environments and often upgrade it every couple of years, to one of becoming experts in a lot of new things.
And when you are dealing with legacy vendors who are just upping the cadence of their releases, without actually providing ways for customers or administrators to implement at the same speed, well then the administrators job isn’t different, its actually hell.
One thing we are doing a lot of research around at the moment is how we might adopt the concept of DevOp. Again it looks like the web 2.0 crowd are coming up with methods and techniques that are making our move from on-premise to cloud more palatable. Ill blog later when I work out where we are going with that one.