Author Archive

I’ve talked in some pre-sales meetings about StorageLink but haven’t really been able to talk about which features work from which SAN vendor.  Citrix has posted an HCL describing just that here.

When xtravirt posted years ago about running esx on vmware workstation, I became a big fan of running virtual ESX, and it was the first time I had my own test esx server.  Well things have come quite aways in the virtual ESX space, now you can run virtual ESX on ESX (link) and HyperViZor are putting info on running an entire “cloud” Lab Manager and SRM on a single box (link).Â
Time to buy more ram.

Datacore recently announced a new product Advanced Site Recovery, basically seems to be enhanced asynchronous SAN-replication with features like 1 click failover/failback which got me thinking about DR and another conversation we had recently with DataDomain.  While virtualization makes disaster recovery dramatically easier, there are still some significant challenges to a great disaster recovery plan.  Basically all approaches boil down to backups or SAN Replication.  Of the two approaches I’ll contrast two of the most interesting approaches, DataDomain on the backup side and VMWare’s Site Recovery Manager which plays on the SAN Replication side.

DataDomain dedupes your backups of your vmdk’s and can replicate them to another DataDomain in a remote site.  VMWare’s SRM uses a certified storage vendors SRA and leverages that storages asynchronous replication to move your vmdk’s to a remote sites, then automates the cutover to separate sites into a few clicks, as well as functionality like testing your dr plan without disrupting production environment.

SRM is a more complete solution that automates a lot of the recovery processes up to re-ip’ing your dr VM’s, but requires the bandwidth to keep up with the incremental block level changes on your SAN.  A T-3 can replicate about 1/2 a TB per day as a general reference point.

Where DataDomain shines is in the lower bandwidth scenarios.  With inline deduplication they can reduce vmdk’s storage on average 40x-60x, I’ve heard of cases of 100x!  They advertising being able to handle 14TB of data across a T-3 in a day!  While recovery is not automated, your dealing with vmdk’s, so its still pretty easy, restore, add to inventory, power on.