Watched some announcements and demos this AM from Microsoft… interesting stuff and much food for thought on direction and ideas. Specifically, you can see a little more clearly now where they see thing going in terms of functionality, performance, data protection, utility, etc.
The Data Amp “event” was a couple of really solid sessions broadcast initially live (and available now on-demand) talking about these things and how Microsoft is approaching them. A few things of note that I was excited to hear about include…
Perhaps not what you think, but for me, it was the database being more intelligent about the data and the ability to perform deep learning and analysis and such on the data in the system, without extracting it to a process. The reason this is interesting is that it preserves all sorts of things – like data security, access controls, compliance, etc.
New term (to me): “At speed” or “At near-human speed”
These were used a few times, but what they’re talking about if I’m understanding (and this is a broad area, so forgive me if I initially get it wrong; I have much to read and learn in this area), is that the database intelligence can be applied in real time, vs. sending information to be massaged in batch.
We’ve talked about big data and velocity and all of that, but this is different. It’s applying things learned from machine learning and modeling in real time. The way I described it was, essentially, as a doctor is entering symptoms and responses from a patient, the database is applying intelligence to those and helping the diagnosis process in real-time. It’s learning from successes and failures and applying that across the users of the system. It’s an incredible opportunity to share learnings in a much more broad manner. The examples they showed included analyzing images and working with other types of data, which is really exciting as well.
Intelligence in the database
This is a fine line. I have to learn more about this and what the difference is between intelligence in the database (as in machine learning) and having a database that intelligence can be programmed into (which we have today with stored procs and .net, etc.).
Deep learning, deep learning, deep learning
Yeah, over and over and over… It’s such an abstract term. It’s like the early days of “big data” and “business intelligence” – we’ll get there, but right now, this is very, very vague.
I was really, genuinely excited about some of the things they were showing – and I have to find ways to talk about it because it’s this weird, emorphous blob of “wow, this is really cool data stuff!” running around in my head. But worth taking a look at, if you have the time.
If not, we’ll keep after it for you.