Buy Percona ServicesBuy Now!

SSD: Free space and write performance

 | July 17, 2010 |  Posted In: Benchmarks


( cross posting from SSD Performance Blog )
In previous post On Benchmarks on SSD, commenter touched another interesting point. Available free space affects write performance on SSD card significantly. The reason is still garbage collector, which operates more efficiently the more free space you have. Again, to read mode on garbage collector and write problem you can check Write amplification wiki page.

To see how performance drops with decreasing free space, let’s run

benchmark with different file sizes.

For test I took FusionIO 320 GB SLC PCIe DUOâ„¢ ioDrive card, with software stripping between two cards, and there if graph how throughput depends on available free space ( the bigger file – the less free space)

The system specification and used scripts you can see on Benchmark Wiki

On graph you can see two line ( yes, there are two lines, even they are almost identical).
First line is result when FusionIO is formatted to use full capacity, and second line is for case when I use additional space reservation ( 25% in this case, that is 240GB available). There is no difference in this case, however additional over-provisioning protects you from overusing space, and keeps performance on corresponding level.

It is clear the maximal throughput strongly depends on available free space.
With 100GiB utilization we have 933.60 MiB/sec,
with 150GiB (half of capacity) 613.48 MiB/sec and
with 200GiB it drops to 354.37 MiB/sec, which is 2.6x times less comparing with 100GiB.

So returning to question how to run proper benchmark, the result significantly depends what percentage of space on card is used, the results for 100GiB file on 160 GB card, will be different from the results for 100GiB file on 320 GB card.

Beside free space, the performance also depends on garbage collector algorithm by itself, and the card from different manufactures will show different results. Some new coming cards make high performance in case with high space utilization as competitive advantage, and I am going to run the same analysis on different cards.

Vadim Tkachenko

Vadim Tkachenko co-founded Percona in 2006 and serves as its Chief Technology Officer. Vadim leads Percona Labs, which focuses on technology research and performance evaluations of Percona’s and third-party products. Percona Labs designs no-gimmick tests of hardware, filesystems, storage engines, and databases that surpass the standard performance and functionality scenario benchmarks. Vadim’s expertise in LAMP performance and multi-threaded programming help optimize MySQL and InnoDB internals to take full advantage of modern hardware. Oracle Corporation and its predecessors have incorporated Vadim’s source code patches into the mainstream MySQL and InnoDB products. He also co-authored the book High Performance MySQL: Optimization, Backups, and Replication 3rd Edition.


  • Fusion IO is well known for it’s bad performance over time. So I would like to invite you to test another SSD drive that use SandForce controller. If you could perform the same test on OCZ RevoDrive it will be great and helpful.

    Here is my reference about Fusion IO bad performance:

  • Haytham,

    I think if you see performance getting worse on FusionIO over time – it is again related to space consumption.
    I for sure will do benchmarks on OCZ RevoDrive as soon as have it on the hands, I do not see it yet available for purchase.

  • Haytham,

    I don’t think the data above shows that Fusion IO has bad performance over time; rather it shows that it has poor performance when *full*, doesn’t it?

    To show performance as a function of time you’d have to fill the card to some constant level (say 50%) and then run it that way for some period of time wouldn’t you?

    Mind you in practise storage fills up over time, so I’d expect most real world use cases to involve progressively less free space as time passes, but that’s not a function of the equipment.

  • if the performance is pretty much the same between space reservation and the 25% reservation why should I wast 25%?
    And are the IO’s affected too?

  • It would be interesting to see, if there is any change in this regard with latest release of FusionIO drivers (2.1 driver + Firmware v5.0.1, rev 42895)


Comments are closed