SQL Server is interesting. If you look at it, it's a database, right? So ultimately the challenge is SQL Server doesn't face it, it's application face it. And the challenge with application faces is, the SQL server is running slow all the time. And so as a Debian developer, I have now challenged somebody else's complaining, and I need to fix things which has nothing to do with what they are actually complaining. And there is the problem.
First of all, I don't know what their challenges are. Second challenge is the SQL itself, because SQL Server is such a brilliant product that there are multiple version of the SQL Server out there. Then we talk with the multiple version of SQL Server, the challenge is now very different. Every product has its own algorithm, and cardinality, or compatibility, which we talk about.
Now when we are running SQL server with the different compatibility level, and the version which we have, the challenge comes out. Like there are times where we don't know what actually is our baseline. That means we do not know what actually we are losing out in terms of SQL Sever performance. This is the challenge. Challenge is, we don't know what we are missing out. Challenge is, there are things which we can get it, we are just not over it. And then the solution is even more difficult to find.
And this is where we need to know a couple of things. There are three ways to address this particular thing. First of all, we need to know the historical trend. Which means like, you can go and take a look at heuristic. Like how each of the version we improve the SQL Server, and what we can get in the future version of SQL Server. That way we can estimate what we are missing out.
Second thing is all about experimenting. Like this, if you look at a product, and if you change 10 different things, sometimes we get one difference. And sometimes we change one thing and see 10 different differences. At that point of time, how do we know what exactly we change, and how to get maximum performance out of it? And that's what we need to know.
And this is where it's very, very difficult to figure it out if we had to do it manually. And that's the time where we have to sometimes depend on a third party tool. Like for example, let me say this way, like Foglight, right. Look at this, spotlight. These are the products which are really going out to help us out. Because they not only show us the trends of what we have done in the past, also can forecast the problem. And also can do the various alerting, and can tell us what exactly we need to focus. What exactly we need to fix, and what exactly we may be missing out.
So yeah sometime, the problem is we don't know what we are missing. But the solution is much simpler if we just know how to look at our history and forecast our future. So that's where we need to look at it, and that's number one challenge. But the solution is not as difficult as we think.
02:52