If this is such a common problem, it’s important to understand the underlying reasons for “database slowdown”, so you can identify the signs as early as possible and respond accordingly. Specifically, you need to know The 3 Enemies of Database Performance, these are the most common issues I’ve seen in my career, and having a handle on them will help you avoid the situation in many cases.
The 3 Enemies of Database Performance
There are 3 big “enemies” of database performance that occur all too frequently. While most of us are familiar with these issues in the traditional RDBMS (Relational Database Management System) arena, they can happen just as easily with a poor data model design or incorrect approach using NoSQL, “NewSQL” or virtually any type of database storage engine. In fact, I recently heard a story from an attendee at a meet-up telling us how a NoSQL engine ran one of his map-reduce jobs for 10 full days! So these 3 enemies can affect us all.
Enemy #1: Table Scans
Let’s start by covering the worst offender when it comes to defeating database performance: Table Scans. A Table Scan will take place when a query requests a row (object, etc.) or more frequently a range of rows (objects) from the DBMS engine. If there is no available index to support the query, or if the DBMS optimizer doesn’t select an available index, then the engine has no choice other than performing a sequential scan of the data set from front to bottom (table, or otherwise).
The Table Scan often accounts for rapid performance degradation – the “hockey stick” graph I showed in last month’s article.
THIS IS A PREVIEW. DOWNLOAD ISSUE 9 TO READ THE FULL ARTICLE.