Clients request our unique blend of creativity and strategy.
Sometimes we even add little umbrellas.
EMC Test Bench
EMC PERFORMANCE TESTBENCH
The EMC Performance Test Bench is an interactive tool that will help you understand the game changing low latency performance that is possible with the DSSD D5 Rack-Scale Flash Appliance. All of the data displayed in this tool
is real data measured from a D5. A user will be able to compare the D5 performance of several random read workloads to that of an SSD installed into the server and against the DRAM installed in the server (the main memory of the
server). What you will see is that the D5 performance approaches that of DRAM and is easily many times faster than an SSD directly attached to the same server connected to the D5.
GOT IT. LET'S GET STARTED.
Walk me through the various tools >
EMC PERFORMANCE TESTBENCH
Select Read Sample Size:
(4K SAMPLES @ 32KB)
EMC PERFORMANCE TESTBENCH
To use the tool:
1) Select a “Read Sample Size” - either 128MB, 256MB or 1GB by clicking the box
2) Review the data on the graph
3) Hide or show the results of all three devices by clicking the “Show” or “Hide” box beneath each device name on the left hand portion of the tool.
4) Learn more about the tool by clicking on one of the help topics.
Select a Read Sample Size
Once you choose a Read Sample Size, we will run that workload several times. Each time we run the Read Sample Size workload we will increase the amount of storage we access on the device we are reading from. This means
each time the device has to search a larger and larger amount of its storage to return the same amount of random data. The device could be an SSD, DRAM or the D5.
We call this increasing area of the disk that we are searching the “Pool Size”. If we only wanted to know how long it took to randomly read 128MB, 256MB or 1GB worth of data from the entire device, we would just have one
data point for each Read Sample Size, for each device. We would in that case be accessing a Pool Size that is equal to the device capacity.
However, this tool will show you more results than just the performance of reading from the entire device. For each Read Sample Size the test will show you the amount of time it takes to read from a Pool Size of 1 GB on an
SSD, DRAM or the DB, and then it will repeat the same Read Sample Size workload again but increase the Pool Size to 2GB, then 4GB, 8GB all the way up to the capacity of the device being tested.
Once you select a read sample size of 128MB, 256MB or 1GB, the tool will automatically plot all of the data points on the graph. You can then hide a given devices data set if you want to just focus on the performance of a
particular device or set of devices.
Reading the Graph
Along the bottom of the graph are the Pool Size values. The Pool Size for each device ranges from “1GB” to the capacity of the device that was tested. Each Pool Size is marked by circle. If you click the circle a note
will pop up displaying the milliseconds required to complete the read samples at that capacity point. There are a maximum of 13 Pool Sizes measured: 1GB, 2GB, 4GB, 8GB, 16GB, 32GB, 64GB, 128GB, 256GB, 512GB, 1TB, 2TB and
4TB. Only the D5 had capacity from 512GB to 4TB.
Along the vertical axis of the graph is the time, in milliseconds, that it took to complete a given Read Sample Size workload, at a given Pool Size.
Lets take the example of running the 128 MB Read Sample Size against the D5. That 128MB test was repeated multiple times at ever increasing Pool Sizes. So you will see several data points in a line very close to the
bottom of the graph – THIS IS REALLY GOOD! The D5 shows very constant, low latency I/O performance even as the amount of storage being read increases. This shows how the D5 fully utilizes all of the storage available in
parallel. We want to see these workload complete as fast as possible, or in other words, plot points closer to the bottom of the graph. 0 milliseconds (no time at all) would be ideal, and moving up the vertical axis
means it’s taking you longer and longer to finish the read test.
In our 128MB Read Sample Size example for the D5, if you were to look at any of the Pool Sizes by clicking on a given circle you will see a nearly constant time to complete the tests!
It gets more interesting if you next select the DRAM data for the same workload. You will see that the DRAM takes a little less time, up to 16 GB (because that is all the DRAM we had to run the test against). DRAM is
expected to be faster than the D5, but as you can see it isn’t THAT much faster. D5 random performance is more like DRAM than an SSD!
Turn on the SSD data for the same workload and you will see that the SSD takes 900 milliseconds – or nearly a second, to complete the Read Samples at various pool sizes.
Why do some data points seem to be missing from the DRAM and SSD?
We have three devices to compare: the DSSD D5, an SSD and the DRAM in the server itself. Although we have 100TB of capacity to use on the D5, we limit the Pool Size on the D5 to 4TB. The SSD we compare to has a maximum
capacity of 256GB, so that is the maximum Pool Size for the SSD. The DRAM we can run our tool against has a maximum capacity of 16 GB, so that is the maximum Pool Size we can run our Read Sample Size test against for
This means that depending on the device that you are evaluating, you will see more or less “Pool Size” data points, because the capacity of the device varies.
A Read Sample Size is the amount of data you’ll read from the device. The 128MB Read Sample Size yields 4,000 random samples, each 32KB. The 256MB Read Sample Size is 8,000 random samples, each also 32KB. The 1GB Read Sample Size is
32,000 random samples, also at 32KB each.