reduce cost

Today EMC introduced a special product in the data protection space which I have been a big fan of over many years – RecoverPoint for VMs.

Having used the traditional RecoverPoint appliances it is all the more great to have data protection at the hypervisor level with all the cool features that RecoverPoint provides and its a very stable product. I was invited to a secret briefing at VMworld yesterday which was in addition to my exclusive conversation with the EMC CTO for Data Mobility Business Unit (RecoverPoint & vPlex). So here is the all you want to know about ‘RecoverPoint for VMs’

It will allow organizations to:

Accelerate Data Protection As A Service For Hybrid Clouds
Empower Data Owners With Application Recoverability & Accessibility
Reduce Protection Cost And Complexity

(click to enlarge)

Powered using the same underlying and hugely popular technology – RecoverPoint for VMs offers local and remote protection. The Virtual Appliance will maintain a journal volume and offer data protection on Virtual Machine basis using a new vCenter plugin.

It is designed for easy and fast deployment – using wizard driven tools (deployment manager) and configuration is very straightforward – especially for those who already run RecoverPoint and are familiar with the interface.

The splitter runs at a hypervisor level and the vCenter plugin discovers the entire infrastructure without requiring a separate UI. All you do is deploy simple OVF file (appliance). The goal is to empower VMware admins to manage data protection, protect critical infrastructure, and orchestrate failover or recovery.

RecoverPoint for VM’s is different than SRM

I would like to provide this level of clarity and I discussed this yesterday with their CTO. Because the splitter runs at Hypervisor level it is easy to manage data protection on a VM basis and thus makes the use of SRM redundant. An added advantage of using RecoverPoint for VMs is that you don’t need to manage data protection at LUN or datastore level. Thus you can mix and match VMs inside a datastore and make maximum use of datastore space. Earlier you had to separate your data protection workload into datastores that could be a part of Consistency group and had to keep other critical VMs out of there if performing application based failover. All those constraints are now relieved and there is great control on the recovery mechanism.

Workflows are wizard driven, protection can be setup across multiple consistency groups or multiple VMware clusters (without requiring separate protection appliances), and write order fidelity can be assured spanning multiple VMs, or across multiple ESXi clusters.

A cool thing is that changes to the ESXi infrastructure can be made on the fly and RecoverPoint for VMs dynamically addresses that. You can Add/Remove VMDK’s from protected VMs and the system will auto provision VMDKs on the target side.

vMotion or Storage vMotion does not affect data protection for the Protected VMs.

Licensing

Since RecoverPoint for VMs is at a VM level the licensing is based on virtual machines. A minimum of 15 VM license needs to be purchased to begin with this software and discounts are available based on volume. There is no transfer of license between traditional RecoverPoint and RecoverPoint for VMs since the software is entirely different.

When you can get your hands on it

EMC will offer trial licenses in October 2014 and I am sure many VMware admins will want to test this out.

For Cloud Providers and those performing Hybrid Cloud computing this is a great solution when it comes to protecting mission critical workloads using RecoverPoint

Some key technical details with a disclaimer – these information below may change down the road and was only made available based on current testing of beta product

A minimum of 2 appliances will have to be deployed per site (source and destination) to ensure high availability and failover
Maximum of 8 appliances can be deployed for larger environments and maximum performance
Site Recovery orchestration built into RecoverPoint for VMs

So I have been intrigued with a lot of cloud computing options and am continuously interesting in reading about what is unique and out there in this area. With the hundreds of ‘cloud’ offerings that are available out there I follow just a select few and was interested to read about EMC’s latest offering. I have for long been a EMC customer and worked with a range of their products and comprehensiveness of the technology platform. The latest news is about the VMAX Cloud Edition that was announced at the VMware Partner Exchange event and it’s a self service access to Enterprise Class Storage. And there is nothing confidential about it.

It is designed to provide mission critical storage services for the public, hybrid, and private cloud. To get into the meat of this solution – they have a self service portal to provision and control storage. Each portal can manage 100Pb of usable capacity across 10 datacenters with 250,000 tenants. This should serve the needs of a lot of large service providers. Multi-tenancy is builtin so logical partitioning can be performed for each tenant and service levels can be changed on the fly to get the right fit. EMC touts their latest offering in response to customer requirements and a detailed explanation is on Chad Sakac’s blog if you are interested. From my perspective the ability to catalog customer/application tiers (mission critical, database, it infrastructure, HPC) into a desired tier within the storage as a service platform is of great benefit. Firstly, you don’t have to worry about the RAID, storage pools, and what not but of bigger benefit is the turnaround time, ability to accelerate deployment, and improve your service level. For e.g. the Diamond tier is set for 1-4ms response times for 3.0IOPS/Gb whereas Bronze tier offers 10-40ms response time with 0.05IOPS/Gb. Data Protection through the use of Time Finder is also of great help because full copies or snaps can be made to protect data at storage level. The product features also include Tenant level metering and chargeback reports prepackaged and available right off the bat. What I am not too sure about and do not like really is the performance related reporting capabilities of various EMC arrays. I am hoping that with a storage as a service platform performance would never be a problem so reporting would not matter that much. Especially, because there is a service level tied with this offering so performance should be as expected.

During blogger briefing EMC also highlighted the availability of Rest API’s to integrate with existing operations and orchestration. I am sure many organizations would use this feature to have seamless continuance of their operational activities and have automated actions associated with them.

Other stuff I would wait for and learn about –

1) How would the Cloud storage environment being offered replicate with a global disaster recovery architecture that apply to many organizations

2) The capability to seamlessly migrate data from internal private cloud to EMC cloud service

3) The ability to use internal tools (that may or may not be API friendly) to work with the cloud service.

I am hopeful and positive that these questions might have been addressed but I will have to find out and that information will be for another post.