Introducing the IBM Storwize V5000 and 7.2.0 Software

ORIGINALLY POSTED 9th October 2013

36,466 views on developerworks

Some of you may have spotted we release another Storwize Family member back in July, but in limited geographies. Today
IBM announced world-wide availability of the IBM Storwize V5000.

As the name suggests, this product sits firmly between the Storwize V3000 and Storwize V7000 ranges, in terms of
scalability, features, performance, and of course cost.

Needless to say the Storwize V5000 runs the same “SVC” software code base as all the other family members, and as I
understand is available to order from today running the current 7.1.0 software versions.

undefined

Storwize V5000 in a nutshell.

The Storwize V5000 is based on a similar enclosure to the V3000 products, and comes in the usual 2U12 or 2U24 control
and expansion enclosures. Up to six expansion enclosures can be attached to a single control enclosure, and up to two
control enclosures can be clustered to provide a maximum of 346 2.5″ drives, an a maximum raw capacity of 672TB using 4TB
drives.

Internally we have a quad core Xeon CPU and up to 16GB of cache per control enclosure. The drives are attached via
a pair of 6Gbit SAS mini-HD connections, giving almost twice the disk bandwidth of the V3000 products.

The build in vs licensed features across the family range are shown below, and as you’d expect the V5000 comes with more
than the V3000 built into the base cost, while not as much as V7000.

Host interfaces can make use of the built in 1Gbit iSCSI, 6Gbit SAS ports, or the optional 8Gbit Fibre Channel or 10Gbit FCoE/iSCSI
host interface card.

undefined



The Storwize V5000 is a really powerful box for the cost, more than capable of saturating the maximum number of
attachable drives, with plenty of horse-power left over for external virtualization if needed.

For more details seehttp://www-03.ibm.com/systems/hk/storage/disk/storwize_v5000/

7.2 Software Announcements

Across the rest of the family, we announce the availablity of version 7.2 software in November. Key features added to 7.2 :

Native IP Replication.

In essence we were so impressed by the performance gains from the Bridgeworks SANSlide appliances that we have supported since June
that we have worked with the team to embed their intelligence into our software stack to enable the same benefits across
our native IP ports. For some time, its always been an item missing from the product family – the ability to replicate without
the need for FCIP routers – so now all Storwize Family products can make use of their native NIC ports for direct
replication.

Here’s the official line :

“The Remote Mirroring function (also referred to as Metro/Global Mirror) is now supported using Ethernet communication links. Storwize Family Software IP replication uses the innovative Bridgeworks SANSlide technology to optimize network bandwidth and utilization. This new function enables the use of lower speed and lower cost networking infrastructure for data replication. Bridgeworks’ SANSlide technology integrated into IBM Storwize Family Software uses artificial intelligence to help optimize network bandwidth utilization and adapt to changing workload and network conditions. This technology can improve remote mirroring network bandwidth utilization up to three times, which may enable clients to deploy a less costly network infrastructure or speed remote replication cycles to enhance disaster recovery effectiveness.”

Global Mirror Increased Throughput

The replication team have been busy, in addition to the native IP form of replication, they’ve been working to streamline
the Global Mirror throughput by removing bottlnecks and in paritcular increasing the performance capabilities at the
secondary site, so that data can be acknowledged back to the primary site as soon as possible. While Global Mirror is
asynchronous, the “SVC” software implementation has always resolved around a minimal RPO objective. The target RPO being
within a couple of seconds. This was enhanced a year or so back with the Change Volume implementation, using FlashCopy to
take copies at both source and target, and co-ordinate the re-sync so you could have RPO’s in minutes rather than seconds.
The net of our short RPO is that the buffering at the primary means you want to secondary to complete as soon as possible.
In the test results I’ve seen, the team has managed to increase single node replication throughput by as much as 4x!

 vSphere API for Storage Awareness (VASA)

The Storwize Family 7.2 software enables users to get more capability out of their VMware environments by being a provider for the vSphere API for Storage Awareness.

SVC Only – Enhanced Stretched Cluster

Prior to this release, stretched cluster configurations did not provide manual failover capability, and data being sent across a long distance link had the potential to be sent twice. The addition of ‘site awareness’ in version 7.2 routes I/O traffic between SVC nodes and storage controllers to optimize the data flow, and it polices I/O traffic during a failure condition to allow for a manual cluster invocation to ensure consistency. The use of stretched cluster continues to follow all the same hardware installation guidelines as previously announced and found in the product documentation. Use of enhanced stretched cluster is optional, and existing stretched cluster configurations will continue to be supported, and can be converted to the new enhanced version with a few simple commands.
    
   
Storwize V3000 and V5000 Only – SAS Data Migration

Data Migration using SAS connectivity on Storwize V3500, V3700, and V5000. Data migration support is standard on all Storwize V3500, V3700, and V5000 systems and this function can now be performed using SAS connectivity to help you easily and nondisruptively migrate data from IBM System Storage DS3200 and DS3500 systems onto Storwize V3500, V3700, and V5000.


So all in all, a great new set of features and functions will be available when we release 7.2 in November. Some other enhancements
include better compression performance – using a new compression algorithm than can be up to 35% more efficient per core, and
better data placement within compressed volumes – to maximise the ability for Easy Tier to determine hot extents. More of this
to come in subsequent posts.

Family Comparison

Finally as promised, here is a handy summary of the features, expansion ability and licenses available across the Storwize Family :

image

One response to “Introducing the IBM Storwize V5000 and 7.2.0 Software”

  1. Hi Barry:
    I would like to ask you for some config options for a V5000. Recently my company has decided to retire a V5000 from production. However the unit is still relatively powerful, being configured with SSDs, 10k SAS and NL-SAS. As this is what I ‘inherited’ I’d like to redo the storage with a different mdisk config. The previous config was done, in a very uneconomical manner, running RAID 1+0 in all disk group.
    For the disk config, it has 9x300GB 10k SAS, 5 x 400GB SSD, 5 x 800GB SSD and 5 x 4TB NL-SAS. My focus is on the SSDs and here are my questions:
    – Can I create an mdisk with mixed drive size? I’m thinking of creating a DRAID6 of 10 SSDs.
    – If the above is doable, can I create volumes out of the remaining space of the extra space in the 800GB SSDs?

    Thanks.
    Bong SF

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: