We hear it all the time, “How does your system compare to your competitor’s system”. That is when the architect or engineer starts to rattle off the different features and how this works and that. But I think you don’t have to go far below the surface to start seeing real differences. We all know the SPC numbers for these systems and try to spin them when our system doesn’t beat the current record. What I do know is there is more to the evaluation of a storage system than just how many bits it can serve in microsecond.
I love cars and all types of cars. This past weekend I got to see one of the best looking, best sounding cars that is made in America. The Chevrolet Corvette ZR1. This car just oozes raw power and from the sidewalk will make men into boys as it rumbles by. But how does it perform? How does it match up with the other ‘great’ cars in that same genre?
Let me introduce you to the SRT Viper GTS. I had to do some research before finding a car so similar that they even look alike. When you look at the hardware the cars are almost identical:
So from the surface you can see these two cars are not only similar in shape but engine size, power and torque. In order to tell which car is better suited, one would need to go head to head and see how these cars perform in certain conditions. Like storage, we compare systems CPU, cache port speed, maximum number drives etc. At first glance you can determine if you are comparing similar systems or not. You may be trying to compare a sports car to mini van.
Now what do you do next? Put the cars on the track and see who is fastest. Drag style. In the storage world, this is done from either a bake-off or looking at benchmarks likes SPC numbers. Now your numbers could show you exactly how your system COULD perform. Now when looking at benchmarks like SPC you have to take in account of how the boxes are configured, how they stack up price wise and even what the green value might be.
But when it comes down to it, it’s all about handling both for cars and storage. As people look at the solutions on paper they may or may not be the right fit. Just because a car can go 0 to 60 in 2.3 seconds doesn’t do you any good if you are driving on the interstate for three hours. But a nice comfortable seat with adaptive cruise control makes your back feel better in the end. Its the things that you sometimes can’t see in the data sheet or someone’s elevator pitch.
IBM does have ways to let you test drive their systems either onsite, remotely or even at a IBM facility. The idea of putting your hands on the equipment and finding the little things that make you more comfortable can effect how you see storage in a different light. If you are in the process of looking at a new storage system, take the time to sit in the car and see how it handles.
Here is a great YouTube video showing the Head2Head comparison of the Viper GTS vs the Corvette ZR1.
Last week IBM published some very interesting results on the Storage Performance Council website. Using the SPC-1 test method, IBM raised more than just a few eye brows. IBM configured 18 200GB SSDs in a mirrored configuration which was able to attain 120k IOPS with less than 5 ms response time even at the 100% load.
IBM used an all SSD array that fit into a single 2U space and mirrored the drives in RAID 1 fashion. These were all put into a pool and 8 volumes were carved out for the AIX server. The previous SPC1 run IBM performed used spinning media and virtualized the systems using the SAN Volume Controller virtualization engine. This gave IBM the top spot in the SPC1 benchmark with over 520,000 IOPS costing a wopping $6.92 per IOP.
This has been compared to the Oracle/Sun ZFS 7420 who published 137k IOPS late last year. When matched to the IBM V7000 results we see the V7000 was $181,029 with roughly $1.50 per IOP compared to the SUN 7420 who came in at $409,933 and $2.99 per IOP. The V7000 was able to perform 88% of the work at 44% of the price. Oracle has not come back with any type of statement but I can only bet they are seething over this and will be in the labs trying to find a way to lower their cost while still performing.
The SPC1 benchmark tests the performance of a storage array doing mainly random I/O in a business environment. It has been called out for not being a very realistic workload as it tends to carter to higher end, cache heavy systems. Now we have a mid range box that not only has the most IOPs trophy but also wins in the ‘Best Bang for your Buck’ category too.
I would have liked to see what the results would have been with the addition of compression to the software feature stack. This going to be a game changer for IBM as the inline compression of block data is way beyond what other vendors are doing. Couple that with the virtualization engine and now I can compress data on anyone’s storage array. The V7000 is definitely becoming a smarter storage solution for a growing storage market.
Great blog articles from both our SVC guru Barry Whyte (no not the singer) and Tony Vandewerdt of Aussie fame toting our recapture of the SPC-1 fastest test results. SAN Volume Controller (SVC) has been pretty much in the top of the finishing results and number 1 for most of its 10 year life. No other virtualization engine out there supports as many platforms or has awesome/ground breaking technology like it.
In fact it was the first real cloud like storage platform as it can move data from one system to the next without causing disruption. I have not seen any other vendor (Netapp, HDS, HP) have anything close to the performance and stability of the system. A couple of things I would like to see change is maybe better snapshot technology and more protocol support but those are not really what this box is designed for.
One of the great facts that Barry points out is this config only had 1920 disks and achieved 270 IOPS per disk. The previous record owned by the HP 3PAR system had the same amount of disks but was only able to squeeze out 450,212 IOPS total, a whopping 69,831 less than the SVC new record of 520,043. The difference is about 15% more throughput.
The SVC’s little brother Storwize V7000 Unified (V7kU) has all of the same software baked in so all that is different is the hardware for the most part. While I don’t think we will see benchmark numbers of 500k IOPS on the V7kU it does have the potential if the hardware could handle it. Maybe IBM will see this and start to think about combining the two products?
Congrats go out to the SVC teams in Tuscon and Hursley for a job well done in deed.