Help get this topic noticed by sharing it on Twitter, Facebook, or email.

General performance characteristics of multi-tenant plans?

I'm a software engineer with experience in traditional self-hosted JEE and Oracle on beefy dedicated server-grade hardware (or the odd Exadata). I'd like to have a rough expectation of the performance of your multi-tenant hosted solution e.g. Pivotal PaaS at the various plans. Are any benchmarks available? I'm not looking for a guaranteed SLA, but a rule of thumb. Something along the following:

Assuming a moderate level of OLTP background activity for the permitted connection pool size...

1. What order of magnitude can I expect roundtrip PK queries to execute? 1/100s of seconds, 1/10s of seconds?

2. What's considered a large table? At what row counts can I expect 1/10 second query execution, subsecond query execution, and < 10 second query execution for:

a. Full table scans, SQL with aggregate functions.
b. Non-unique index scans.
c. Non-unique index scans, SQL with aggregate functions.
d. Unique key / PK queries.
1 person has
this question
  • Hello,

    Unfortunately I can't give you hard benchmarks for your questions. On a shared database environment load can vary widely and unpredictably; you have many different customers all accessing the system on their own schedules.

    In general, under normal load conditions, query execution time is closer to tens of milliseconds rather than hundreds but there are many factors which can affect that, as you know.

    If your driver is replicable performance, then our mission-critical product line - where you have exclusive access to the resources of a cluster - is probably a better fit for you.

    If you can direct your questions to they can probably answer in more detail than I have at hand.
  • (some HTML allowed)
    How does this make you feel?
    Add Image

    e.g. sad, anxious, confused, frustrated kidding, amused, unsure, silly happy, confident, thankful, excited indifferent, undecided, unconcerned