Benchmark Documentation

For the interested reader we provide here the full documentation about our benchmarks.

Cluster benchmark

Replication benchmark

Single node benchmark

22 million transactional updates/s: DVTDS achieves unprecedented level of LDAP performance

What if you run a global business with hundreds of millions or even a billion users? Facing them producing a million updates per second in world wide distributed data base. Or ten million. Or more. And changing content that must be visible to others in real time. Do you buy or lease tons of hardware, rack space, air condition, energy, cabling, security staff, administrative staff, in one word: Dinosaur footprint?

What if you run a particle accelerator or a wind channel generating vast amounts of experimental data within milli seconds for every single shot or try? The same questions apply and at the end of the day it all boils down to operating costs. Even if you definitely can afford more hardware and have scalable data base software the scaled environment as a whole becomes a problem. With ever more nodes the setup rapidly approaches a point where pure quantity turns into deteriorating quality and decreasing maintainability.

DVTDS is engineered for these challenges with no compromise. Through the entire development process we always ask if this or that could be made faster or need less memory or both. This is not just doing an engineer's job. This is something in the middle of dedication and obsession. We are out to create the most powerful directory server in the world and it seems that we are way down the road. In terms of performance it already compares favorably to any other data base, including NoSQL key value stores. But with its support for complex data models, online schema management, view tables, triggers, access control and attribute level history it is really outstanding. In particular we showed: