Tuesday 30 October 2012

Hitachi Unified Storage Specification and Features


Hitachi Unified Storage Specification and Features

Few days back I attended the Hitachi HUS modular training, and while the training was going on, as I have  already worked on NetApp storage, so I was keep on comparing Hitachi each feature with the NetApp unified storage features, but apart from the architecture (hardware & CPU and cache) I have not found anything new in Hitachi unified storage. While one thing I can say that installing the Hitachi storage and upgrading the Hitachi firmware’s are quite easier than the NetApp storage and even managing is also quite easier than the NetApp GUI.

Hitachi unified storage is not unified in Hardware , but it is unified from software front, means they have Hitachi Command suite  for management of Hitachi SAN/NAS and even the other VSP, ASM, others product.

When I asked that why they have not merged all the hardware in single hardware like NetApp, their straight answer was they don’t want to compromise on performance, more hardware means good performance, because each hardware have its own memory, ram, CPU to perform its own task, while merging in single hardware the load on central CPU will increase and then there will be more hardware failure, decrease in performance.

Well I have not used any Hitachi product so I cannot say that whether they are good or not in performance but the examples of some of their customer  feedback they presented in training, that quite say that Hitachi is really good performance storage box. If you see their controller architecture their you can see that there is 6GB/s connectivity between the its RAID Processor (DCTL) chip or we can say that the between the two controller there is 6GB/s  connectivity which help their controller to do the load balancing, that’s good, because this was new to me, because I have not seen this in NetApp, yes controller failover and failback happens in NetApp , but NetApp never say that their failover /failback link do any type of load balancing activity.
Yes I know that in NetApp the disk are owned by the respective controller and their load will be handled by that owned controller only, and each controller have some load handling capacity, but lot of customer don’t understand that and they keep on adding the disk shelf to one of the controller in the hope that in future they will add to the other controller and by this one of the controller which have more number of disk owned by the others have to handle lot  more iops than other and hence the utilization of that controller increases and the performance get decreases.
As I have also worked in some of the EMC VNX series storage box, which are unified storage box of EMC, well there they strictly recommend to keep adding the expansion shelf alternatively, for ex if one expansion shelf is attached to the one of the controller bus then other expansion shelf should be attached to the other controller bus, for load balancing.
So that Cleary state that neither the NetApp nor the EMC have such type of 6GB/s internal connectivity between the controller which can do the automatic load balancing like Hitachi, but still I cannot write much because I  do not have any experience in Hitachi automatic load balancing features so could not say that whether it really works fine .
But by doing some of enquiry with some of the colleagues there who have good experience in Hitachi storage stated that they never seen much of hardware failure in Hitachi except the disk, I mean like controller failure, PCI card failure, or fan failure or power supply failure types. But I can say the hardware failure in NetApp is quite higher than the other vendor storage product. That’s my personal experience and I don’t know why, even you can experience that suddenly some mail will get dropped in your inbox from NetApp stating that please do this or that firmware upgrade urgent to save you controller getting failed (W.T.F) and then you have to plan for that activity, so bad, my personal experience, NetApp have lots of bugs and the good part is they keep working on them.
While in Hitachi training they were more focusing more on the cache part that means in Hitachi storage you can do lot of cache modification, like you can set the cache size as per your application and the data is first written to the cache and then to the disk, so that if there is any disk failure the cache data will get copied to the flash memory and the content can remain there for infinite time (means until you get recovered from your power failure).
Even there is one more features in Hitachi unified storage that you can fix size of the block of greater size, like every storage divides the data into their some default block size like 4kb or etc. and then store it in disk, but in Hitachi you can increase the block striping size of greater size also, but these all activity should only be done when you get the recommendation from the Hitachi technical team.
Now striping the block of bigger size is really a fantastic feature and it really improves the performance, so by all this good features you can see that Hitachi is more concentrated on performance and that really good and best things of Hitachi storage apart from this basic installation of Hitachi is quite easier compared to NetApp and EMC and then the part replacement is also quite easier than the NetApp and EMC , well for the part replacement you don’t to apply your extra brain just go and remove the faulty part that’s all .
But still if you see that today’s storage world is talking about deduplication and automatic tearing, which is still not there in Hitachi storage, and these features will be soon launched in coming version. 

Below are the file module specifications:

Below is HUS block storage specification:
Now in HUS unified storage you have to buy the block level storage and file level then only it is unified, so when you say that u wan Hitachi unified you will get block and file both.


Well I have not deeply explained the features of Hitachi storage, because still I need to work in this product and by working on any product you can easily understand about it, so in future I will be writing some more blogs on Hitachi product.