SAN Selection

What SANs would you recommend for the following

- 20Tb Usable data
- VM infrastructure / integration
- HP Blade server / chassis (current)
- Commvault backup
- replication in real time to 2nd SAN at DR site on dark fibre
- NFS / CIFS shares
- Thin provisioning

We are looking at

IBM V7000
HP P6300
HDS AMS 2100
Netapp FAS3200

Any comments positive / negative would be helpful....
Clazman-ITAsked:
Who is Participating?
 
Paul SolovyovskyConnect With a Mentor Senior IT AdvisorCommented:
I would also take a look at the EMC VNX product as well.  

I have worked with the EVA and the Netapp out of the list provided so I'll give you what I know
.
- 20Tb Usable data

Both can support 20TB
HP: Strictly block so you'll use a good bit of space, snapshots are copy on write so take more room then the Netapp.  Now offer thin provisioning
Netapp:  Better use of space, HP is monolithic in storage management, not too much you can do to increase space (no dedupplication, etc..)

- VM infrastructure / integration

HP:  The vCenter plug-in in very basic, not too much you can do with it other than monitor and provision. Still not there yet IMHO
Netapp:  Integration allows you to: provision, monitor, backup VMs, etc..much better then HP (using vmware snapshots as part of backups is huge to quiesce the VMs)


- HP Blade server / chassis (current)

You can use the blade chassis to connect to either Netapp or HP without issue

- Commvault backup

-Suppprts VMWare, agent backups, and NDMP (For Netapp and EMC)

- replication in real time to 2nd SAN at DR site on dark fibre

HP:  Contiunous Access needs MPX controllers and if you're using a blade chassis you'll need another FC switch in addition to MPX controllers for FCOE access for replication to work.

Netapp:  All you need is a licensse and provides throttling and compression for replication


- NFS / CIFS shares

HP:  Does not support
Netapp: This is Netapp's sweet spot, it offers NFS and CIFS and with NFS you can run your VMs with deduplication on the NFS (or iSCSI) with 30-80% deduplication on basic virtual machines.  CIFS - the Netapp acts like a windows server and can be managed with Windows Computer Manager

- Thin provisioning

HP and Netapp provide thin provisioninnig

I would also add deduplication, Netapp deduplication is superior, typically 50% CIFS, 30-50% on VMs so actual storage requirements much more efficient.
Netapp can also do FC/ISCSI/NFS/CIFS/FCOE, while HP is noly iSCSI/FC

If you have a C7000 chassis you can put the HP 10GB switches and get really good use of the Netapp for high iops environment.

Currently doing a project with HP 6500 and Netapp 6020 and got a handle of both sides..
0
 
tomagoCommented:
An advantage to the HP (former Lefthand products) is the ability to use a virtual SAN to replicate to.  This means you can setup a super cheap storage array to replicate to.   This eliminates the need to purchase 2 SANs so you can get a much better SAN for production and just get cheap storage for replication.  

This also is a major advantage when bringing in smaller branch office since the cost is so low.  You can still put in a virtual SAN (for little over the cost of a Drobo or other SMB SAN) and manage it centrally with the rest of the Lefthand nodes.  
0
 
Paul SolovyovskySenior IT AdvisorCommented:

@tomago

The HP P6300 is the EVA line, HP finally decided to simply their naming from MSA (now P2000 DotHill rebrand, P4000, Lefthand, and EVA)

P4000 is strictly iscsi which negates the HP Blade Chassis FC capability if the need arises, no NFS or CIFS.  It also doesn't eliminate the need for 2 SANs as it needs to be replicated to a DR site. What it does provide is loadbalancing and failover but so do other systems with multiple active/active controllers
0
Improve Your Query Performance Tuning

In this FREE six-day email course, you'll learn from Janis Griffin, Database Performance Evangelist. She'll teach 12 steps that you can use to optimize your queries as much as possible and see measurable results in your work. Get started today!

 
agilerayCommented:
We are a Hitachi House, but found that difficulty with upgrades and pricing. Hence moved to P6300.

Basically we only use Fibre Channel technology for host connectivity, because it is most mature and least problematic.

The P6300 is a very easy to use system and interface and the disk group and lun level raid technologies are unique to HP. Now they have added the thin provisioning the lun migration from one raid type to another it has made the decision even more easier.

Remember there is no substitute for spindle count. More disks you have the faster you can get to you data (generally).
0
 
Paul SolovyovskySenior IT AdvisorCommented:
@agileray:  I agree with some points you make but the need for replication: MPX controllers and added FC ports (especially when you've used up the FC ports in the blade chassis for your blades) can get quite expensive.  
0
 
andyalderCommented:
Another option from HP - the 3PAR series, now P10000. Main advantage I see is the number of host ports allowing you to do away with fibre channel switches (although not in a blade environment unless you use pass-thru modules).
0
 
Clazman-ITAuthor Commented:
Some great comments thanks...... we know the Netapp is the highest priced model we have looked at... but our supplier has told us that the 3par is overpriced
0
 
Clazman-ITAuthor Commented:
sorry over priced is the wrong word...should have said..... well into the next price bracket :-)
0
 
agilerayCommented:
@paulsolov - Agreed that it is cheaper with iSCSI / FCIP to do replication etc.

But we let the array be just a block storage, as it will never be at the level of understanding data or have the ability to index it.

We use backup / replication software for our data movement. This makes it much easier to bind with our application and solutions.

But I guess, each to their own. And there is no absolute right or wrong way of doing it.
0
 
kevinhsiehCommented:
I would also look at what Dell has. I can tell you about the EqualLogic line as I have been a customer for 4 years, and they also have Compellent now.

The Dell EqualLocic is on the leading edge of supporting VMware virtual storage API for offloading storage commands like copy and write zero to the array. It will also support reclaimation of thin provisioned space once hypervisors/operating systems have a way to do that. It has thin provisioning and thin clones. Replication over TCP is built-in and included in the purchase. All features are included in the purchase price, so you never have to worry about writing a check to add a new feature or expand a license count; just keep active maintenance. Maintenance is very affordable, and they don't try to get you to rip and replace old equipment like some other vendors do. They fully support every unit ever made with the current firmware, and all of the models can be run together as a single group.  

If you really want CIFS/NFS they do have an add on controller FS7500 to do that. It is the first time a new feature requires an actual purchase.

20 TB is more storage than what you can generally get in a single shelf of disks. If you have 2 or more units, EqualLogic can automagically tier your data among the units so that your most active data is on the highest performing tier. There is a new 24 drive PS6100E with raw capacity of 12/24/48/72 TB using nearline SAS drives of 500GB-3TB sizing. There is also the PS65x0x with 48 1 or 2 TB SATA drives, or 48 600 GB 10K SAS drives. Those are some pretty beefy units with lots of IOPS to power you environment. I expext that the PS65x0 line will get refreshed with 3 TB 7.2K drives (SAS or SATA) and 900 GB 10K SAS drives pretty soon, since those are now available on the PS4100 and PS6100 lines.  
0
 
kevinhsiehCommented:
I would also look at what Dell has. I can tell you about the EqualLogic line as I have been a customer for 4 years, and they also have Compellent now.

The Dell EqualLocic is on the leading edge of supporting VMware virtual storage API for offloading storage commands like copy and write zero to the array. It will also support reclaimation of thin provisioned space once hypervisors/operating systems have a way to do that. It has thin provisioning and thin clones. Replication over TCP is built-in and included in the purchase. All features are included in the purchase price, so you never have to worry about writing a check to add a new feature or expand a license count; just keep active maintenance. Maintenance is very affordable, and they don't try to get you to rip and replace old equipment like some other vendors do. They fully support every unit ever made with the current firmware, and all of the models can be run together as a single group.  

If you really want CIFS/NFS they do have an add on controller FS7500 to do that. It is the first time a new feature requires an actual purchase.

20 TB is more storage than what you can generally get in a single shelf of disks. If you have 2 or more units, EqualLogic can automagically tier your data among the units so that your most active data is on the highest performing tier. There is a new 24 drive PS6100E with raw capacity of 12/24/48/72 TB using nearline SAS drives of 500GB-3TB sizing. There is also the PS65x0x with 48 1 or 2 TB SATA drives, or 48 600 GB 10K SAS drives. Those are some pretty beefy units with lots of IOPS to power you environment. I expext that the PS65x0 line will get refreshed with 3 TB 7.2K drives (SAS or SATA) and 900 GB 10K SAS drives pretty soon, since those are now available on the PS4100 and PS6100 lines.  
0
 
madunixCommented:
One of the project, a company need to have mission critical data available and consistent in all three sites (10 TB Data storage), so we implemented a three-site business continuity strategy based on IBM System Storage DS8000 as long-term solution, the SAN's are attached to open systems (Windows, Linux, AIX).

The client has a primary Production (A) and a secondary Resumption (B) data center 10Km (DarkFiber Connectivity) plus a third data center (C) DR approximately 150 kilometers away (MPLS 100Mbps connectivity). The following licensed functions are used Metro Global Mirror license, Metro Mirror license and Global Mirror license.


http://publib.boulder.ibm.com/infocenter/dsichelp/ds8000ic/index.jsp?topic=%2Fcom.ibm.storage.ssic.help.doc%2Ff2c_modelovervw_1w2zlp.html
http://www.ibmsystemsmag.com/aix/storage/software/Disaster-Recovery-x-3/?ht=
http://www.redbooks.ibm.com/redbooks/pdfs/sg246786.pdf


The main challenge in SAN-SAN replication is to have enough bandwidth (media link), for replication is Dark Fiber would be the best for SAN replication.
0
 
madunixCommented:
Regarding CommVault.....

CommVault is aimed at Small/Medium/Large companies and is feature packed. You can run backup, archive, content indexing, single instancing, and a list of other features from one console.

Commvault is relatively inexpensive, but you also need to factor in the cost of re-training and the
3 to 6 month time that it will take your backup admins to get used to a new product. You need to consider what your needs are and your current infrastructure and future growth rates, as you know
Backup products need to be sized and installed correctly.

A good place to start is to either contact an independent consultancy to help decide Proof of concepts are also a good way to help with the decision which vendors are more than happy to help with.

Here are some case studies (May have some comparisons between the different products):
http://www.commvault.com/products-backup-recovery.html#t-2




 
0
 
Clazman-ITAuthor Commented:
None
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.