What is unified storage?

A single platform support multi-protocol such as FC, ISCSI, NFS, CIFS, FCoE and etc. It also support file, volume and block level data access.

You will able to get the unified box from either Netapp, EMC and HDS in the market now.

Well, seems like all the major storage players are aligned and agreed with the Unified Storage architecture now. Let’s share some of my personal experiences here. For my pass experience as an end users, I had met up with numbers of Storage experts from different principle. I often received those comments as Unified storage will slow down the performance, not efficient and they are not block base solution from multiple experts. No right or wrong for what they had commented, but with the storage trend today, they may need to reconsider the statements they had put in front to the customers previously.

Personal opinion, the storage transformation majority are caused by Virtualization and cloud computing as you could easily decide the protocol or methodology for different deployment, which had forced the storage vendor to re-consider their approach for SAN offer to support unified protocol in single platform.

As I know EMC have EMC Storage Viewer, EMC Celera SRM Failback Tool & EMC Celera SRM Failback Tool and NetApp have Rapid Cloning Utility (RCU), Virtual Storage Console (VSC), SnapManager® for Virtual Infrastructure & SANScreen® VMInsight vCenter Plugin. I’m not going to discuss about the vCenter plugin features here.

Does vSphere admin really need VMware vCenter Storage Plugin?

For most of the company, they have many IT teams such as network team, storage team, server team or vSphere team. Imagine if you were part of the vSphere or Storage admin team, do you think vSphere admin team should overwrite storage guy job? Frankly speaking, these kinds of politics already happen in many companies today. VMware did a good job by having VAAI API enable. The storage guys should focus on their storage and vSphere admin should focus on their VM environment.

Personally I’m not familiar with most of the VMware vCenter Storage Plugin. Do you think VMware vCenter Storage related plugins is a must? What is the benefit or advantage? Do you think storage vendor should keep working on this? What do you think?

General de-duplication solution:

1. Primary SAN de-duplication
2. Appliance target base de-duplication
3. Host base de-duplication
4. Software base de-duplication

I may had missed out some of them, but these should cover majority of the de-deduplication solution in the market today.

Primary SAN de-duplication allow dedupe happens on the primary SAN,  where most of the active data were stored. This should be the highest cost / GB that running in your environment.  Primary SAN de-dupe happen on Netapp and EMC storage today. While both claim they are doing de-duplication, you may need to take note the block base de-dedupe and file base de-dupe are different approach.

Primary SAN de-dupe allow you to achieve disk space saving up to 50% or more, especially in virtualization environment. As we all are familiar with the virtual machine deployment, most of the virtual machines are usually cloned or deploy from the standard template, which will contains numbers of duplicated block data on the primary storage. De-duplication allow administrator to re-claim more space from the primary SAN, and reduce the TCO of overall virtualization strategy. I will suggest de-dupe to be enable for virtualization deployment and exclude the volume or LUN that require high I/O performance especially for Databases.

Appliance base de-deuplication – example Data Domain VTL, Quantum VTL and etc. These solution will be more specified to backup and recovery. As an example, it allows users to perform daily full backup with minimal disk space consumption with de-deduplication on virtual tape library. This is not on primary SAN as your VTL are not the primary SAN where the data are stored

Host base de-duplication – Avamar. This is another great solution in the market; allow data to be de-dupe before process over for local or remote backup. As an example, it will allow users to minimize the bandwidth requirement for remote office or location, and centralize manage the backup and recovery on primary data center.

Software base de-duplication which allows users to convert direct attach, NAS, tape or SAN attach storage, to be target base deduplication machine. This is heavily relying on the management software to perform the data de-duplication and not the hardware appliance itself.

No right or wrong, just to share the different in general how you should consider about de-duplication solution today in the market. Each of them has the unique position in the market for different perspective.

1. Higher consolidation ratio in virtualization with the latest CPU, memory and servers performance
2. Higher capacity disk drive which significant reduce the number of storage expansion require
3. Virtual I/O which allow the higher utilization of the available bandwidth in data center
4. UCS to provide consolidation and higher capacity of bandwidth to support network and storage utilization in the data center

I may had not included all the facts here, but these should be good enough to point out the changing happening to the current market. I had been impressed with the IBM 3950 M2 previously but now, the Unisys alliances with VMware is another choice which provide similar features for the large environment which contains large number of machines running. The unisys Enterprise Server ES7000 could be the ideal machine in this case. 16 physical cpu sockets which contain 6 core each from Intel Xeon 7400 series of processor, and maximum of 1TB of memory with 8GB memory module in a single server. If we do a calculation, you should able to have 96 cores of CPUs in a single server. This will allow more virtual machine to be consolidated into it.

You can view here for more information about the hardware details.

So, with all the technology available, we should able to have our data center running with less number of server racks for servers, storage and networks. This had significant reduce the space require to run a data center, and allow the users to expand the life cycle of the data center in a longer term. If I have to choose to invest 1 Million on DC expansion VS 1 Million of Virtualization, servers and storage, I will definitely choose the 2nd option which provide additional capacity and reduce the space require by a data center. In a longer term, the real estate cost, power & cooling always cost us a big sum of dollars to keep the data center running.

In the near future, there will be 8 cores and 12 cores CPU in the future market. When it come to the time for us to refresh the current ESX servers which run on quad cores or 6 cores base server, the new servers capacity may be double the current servers that we using, and for sure it will allow more machine to be consolidated.

Equipment List

1. 2 x R710 with Intel 5530
2. 2 x PowerConnect 5424
3. MD3000i with 15 SATA Disks
4. Software ISCSI initiator from vSphere 4

I was impressed with the functionality of the Dell PowerConnect switch which provide the important features as link aggregation, jumbo frames, ISCSI optimization and etc. Here is what I did. I had enable the jumbo frame on the specify port that connected for the storage and ISCSI connectiones. At the same time, I had also turned on the MTU configuration and jumbo frame setting from the MD3000i management console. You may need to repeat the same action for every data ports that provided by the storage controller on MD3000i. Enable the Jumbo frame and configure the MTU value of 9000 on it. On the PowerConnect switch,I had configured the specify ports on ISCSI optimization and jumbo frames enable. Now it come to the vSphere level, which I had to manually enable the ISCSI vswitch to support the jumbo frame and put in the right value for the MTU.

Command for reference

#esxcfg-vswitch -l    
This will allow you to list all the virtual switch you have on the ESX server

#esxcfg-vswitch -m 9000 vswitch1  
this command is to enable the mtu value of 9000 for every nics that connected to the same virtual switch which provide the storage access to ISCSI

#vmkping -s 9000 192.x.x.x   
test the jumbo frame setting

The outcome of the jumbo frame enable, I had able to achieve higher throughput on the MD3000i I had and I am happy with the performance of the storage.

Select Change datastore at the 1st screen. Note that the option 1 is use for Vmotion on virtual machines. Option 3 is only available when VM is been powered off

Select the right datastore you wish to move your VM and click next to continue

Validation success and choose the option you need from here. 1st option will remain the original configuration of the VM. 2nd option will go for thin provisioned format. 3rd option will be thick format for full allocation of the virtual machine. Press next and you are ready to go for storage Vmotion.

This features is excellent as it simplify the storage vmotion process to the administrators.

Click here to read more about the Best Practices for SQL Server on VMware

To share my experience here, I had successed to virtualize numbers of SQL in my environment. What I had done is similar to his suggestion and our SQL server is performed well as it needed. I will encourage anyone who are planning to virtualize their SQL servers to read this post, and it should be helpful for your virtualize plan.

Personally, I had 3 of it running in my environment now and of course, I had spent a huge number of times to really test out all the solution and compare apple to apple to further study of my requirement. Here are some reading I get from my test for 3 solution above.

For I/O performance, I am using the IO meter as to generate the max output to the storage. I manage to get 190MB/s for my FC storage on 100 % read, and I only managed to get the output of 170MB/s on ISCSI, but on NFS, I only managed to get at 125MB/s. If we look at the number here, and compare to my high load R900 ESX host, actually it is more than enough to handle the number of VM I have in the ESX host. There are a lot of article introduce NFS, and the features of Netapps, but remember, ESX itself had improved from time to time. Thin Provisioning is no longer new thing, as you may get it with many vendor today by free as well as volume snap. But Dedupe will still be an interesting technology in netapps. If we look at FC, you will mostly lose all this technology, but it provide a stable performance as you need to virtualize some heavy environment moving forward. I am currently still test out the storage solution before I decide to go with a final choice. I will suggest to keep an eye on the technology trend as we may also remember that the VMware ESX 4 is on the way, which may change the game of storage strategy significantly.

Click here to enlarge the picture

Attach is the screen shot which is been captured during my test.  Our ESX server is connected with 2Gb fiber channel dual connection from the ESX host to the Fiber Switches in our environment. If we do a calculation, 190MB/s will be 1520 Mb/s which is 76 % of the 2Gbps FC connection utilization. In between, there will always be some latency lost from the host level to our FC storage. My SAN storage is currently 75% populated with disks in active, this performance is consider good enough for my personal point of view. Bear in mind, this is only utilize 1 physical FC connection from my ESX host. Another interesting we found, that my R900 with more than 35 VMs per host, in the peak hour, it will not go more than 50MB/s. In this case, it seems that our environment is still under utilize from the I/O perspectives. Now it came to the point that, we had invested so much money on the fiber switches but we did not fully utilize the resources we have. Imagine that each ESX server will consume 2 HBA connection, and we only utilize less than 50% of the available bandwidth. The next move, I will definitely looking for something to reduce the FC ports utilization on our VM farm and data center as well. I know Xsigo and Cisco both have their idea to do this. You may need to have a look on VP780 from Xsigo or Cisco Nexus 5000 or 7000 for the next generation data center switches.