Single-Socket OLTP Systems?

Joe Chang tosses a hardware-related bomb:

Today, it is time to consider the astonishing next step, that a single socket system is the best choice for a transaction processing systems. First, with proper database architecture and tuning, 12 or so physical cores should be more than sufficient for a very large majority of requirements.

We should also factor in that the second generation hyper-threading (two logical processors per physical core) from Nehalem on has almost linear scaling in transactional workloads (heavy on index seeks involving few rows). This is very different from the first generation HT in Willamette and Northwood which was problematic, and the improved first generation in Prescott which was somewhat better in the positive aspects, and had fewer negatives.

Joe knows a lot more about this than I do, but I’m very hesitant about this for two reasons.  First, scale.  When we start looking at hundreds of concurrent requests, would a single-socket machine really work?  I don’t know to answer to that, but in my simplistic “more is better than fewer” rule of thumb, I’d err on the side of caution, especially if it isn’t my money paying for this.

Second, there are batch processes and large background activities which occur even on extremely transactional OLTP systems.  Think about running CHECKDB or ETL processing or troubleshooting/monitoring processes.  These are going to be processes which benefit from parallelism, and if you’re seriously limiting core counts (which a single socket would necessarily do), you might end up in a bad way when they run even if your “normal” workload performs a little better.

Related Posts

Avoid Scalar Functions In Computed Columns

Daniel Hutmacher shows why you should not include scalar functions inside computed column definitions: Scalar functions can be a real headache when you’re performance tuning. For one, they don’t parallelize. In fact, if you use a scalar function in a computed column, it will prevent any query that uses that table from going parallel – even if you […]

Read More

Azure And The Kappa Architecture

Jared Zagelbaum describes the Kappa architecture and points out how there’s limited built-in support in Azure for it: You can’t support kappa architecture using native cloud services. Cloud providers, including Azure, didn’t design streaming services with kappa in mind. The cost of running streams with TTL greater than 24 hours is more expensive, and generally, […]

Read More

Categories

April 2016
MTWTFSS
« Mar May »
 123
45678910
11121314151617
18192021222324
252627282930