- Support, it is superb!
- InfoSight, especially VM Vision.
- Easy of setup and management, no need for storage experts.
- Performance, just delivers on all that it promises.
- Maintenance price is fixed for 5 years.
We do not need to employ storage specialists to set up and manage our storage.
InfoSight gives us in-depth knowledge to ensure our storage and associated systems are performing well, which is more than we could even get from previous storage systems.
Support for near synchronous replication.
Currently the smallest snapshot window is 15 minutes, that means if we replicate data from one array to another array we will be up to 15 minutes out of date, this is the RPO.
We have a small number of systems that need to be replicated in such a way that we do not lose any data if we need to fail-over to the replica on the second SAN, or at least it should be only seconds behind.
Some SANs, like NetApp, have a synchronous replication feature, so when data is written to SAN1 a copy is sent to SAN2, only when this has been written successfully at SAN2 does the data on SAN1 get flagged as written. This means both SANs are in-step.
This is great if you have the SANs locally and have a fast connection, but if they are in different data centres and the WAN link is slow you get latency issues while the data is written and acknowledged by SAN2.
When I talk about near synchronous replication I am thinking of how DoubleTake from Vision Solutions works, I have been using this for over 15 years to do this type of replication.
This works by copying the block of data on SAN1 directly to the local volume, but taking a copy and putting it into a buffer. This means latency at SAN1 is not compromised.
The copied block is then sent across the WAN to the second SAN, where it is written and acknowledged back to the DoubleTake system on SAN1. Once the acknowledgement is back the local copy on SAN1 is deleted.
This means if any latency on the WAN does not impact the speed of storage on SAN1, although the data on SAN2 could be a little bit behind, but in most cases we are talking seconds at the mots, and once any congestion on then WAN link has gone the replication is nearly as fast as a pure synchronous system will achieve, so for most purposes it is as close to real-time as required.
If the WAN goes down, or you need to take SAN2 off-line for any reason, the only impact on SAN1 is the buffer grows with data to be replicated, and once the WAN or SAN2 is back on line this is flushed to SAN2 as fast as the WAN can transfer the data.
You need a ‘re-synch’ feature in case the volume on SAN2 is corrupted, this checks the CRC on the data blocks between the two systems and re-transmits any blocks that do not match.
No, none at all.
The CS210 is very limited and could not be expanded despite having spare slot’s for drives. But we were able to scale-out two of these to create a larger single unit.
But we have first generation hardware that cannot be upgraded unlike newer systems.
Amazing. Nothing is too much trouble, and they have a wealth of information they can access to help you.
Biggest problem was fitting the arrays into the racks – need to bend the rails (shoe or hammer required). Fixed in the latest models.
Took less than 30 minutes to get up and running and connected to our servers.
Simple – one price for the equipment, one price for the support and maintenance (fixed for 5 years), any updates and improvements and all licences included.
Looked at options from Hitachi, EMC, NetApp and HP.
Make full use of their Pre-Sales engineers, make them your best friend, they will stay with you for the journey and can assist you get the best out of the equipment.
Don’t be frightened of asking Support for help, if you authorise full data collection they can look at a wealth of information on your behalf covering more than the actual storage elements.