There’s a lot of talking at VMworld this year about Vvol but if we move one step back from the buzzword and try to understand the “why”, what do we see?
we see that VMware wants to bring the data services to be applied at the VM level, things like cloning & snaphsots, replication etc, to take place at the VM level as oppose to the lun (VMFS) level.
what if I were to tell you that it can be done, TODAY?
Enter RecoverPoint For Virtual Machines A.K.A “Project Mercury”
EMC looks at data protection as a continuum that spans protection tiers to provide everything from continuous availability to replication to backup and archive. These tiers all need to work together in a complimentary fashion, creating a continuum that is aligned to the most stringent requirements for zero downtime and no data loss, to various point in time copies, to backups and to secure long term retention with archiving. Only EMC can deliver this continuum.
Today, we will focus our discussion on the left-hand side of the continuum, CDP and remote replication for operational and disaster recovery use cases.
EMC is proud to introduce the newest member of the EMC RecoverPoint family, RecoverPoint for Virtual Machines, a simple, more efficient and agile data protection solution for today’s and tomorrow’s applications in VMware-based cloud environments. This release provides local and remote replication in combination with continuous data protection for per-VM recovery to any point in time, with built-in automated disaster recovery orchestration. RecoverPoint for Virtual Machines brings the enterprise quality and advanced features of EMC RecoverPoint, with over 20,000 units installed worldwide, to a 100% software offering. RecoverPoint for VMs will:
•Accelerate data protection as a service for private, public and hybrid clouds for both enterprises and Service Providers. The fully virtualized and VMware-integrated DP solution comes with a robust set of APIs, enabling our customers, SPs, and technology partners to build secure cloud service offerings for zero and near zero RPO for virtualized mission critical applications. Additional integrations into EMC’s Hybrid Cloud Solution and VMware’s vCD next year will complete the vision of delivering DPaaS for hybrid clouds.
•Empowers VMware administrators and application owners to discover, configure, manage and monitor VM provisioning and control their own protection policies and automated recovery, ensuring that VM service levels are met.
•Reduce protection cost and complexity through an easy to install (existing ESXi servers) and manage (thru vCenter) storage agnostic (supports SAN/vSAN/NAS/DAS) DP solution with built-in automated DR orchestration (no additional SW) and superior WAN optimization capabilities, thus significantly reducing operational costs.
The RecoverPoint product family is comprised of two different products targeted at two different audiences, solving different business problems.
•RecoverPoint is targeted at the storage admin and protects storage LUNs, is managed thru EMC Unisphere (a storage-oriented interface), is deployed on physical hardware appliances, using embedded storage array splitters in VMAX, VNX, and VPLEX and supports over 50 storage systems, including EMC and 3rd party arrays using the VPLEX splitter.
•RecoverPoint Virtual Edition for VNX is the virtual substantiation of RecoverPoint and a more flexible deployment option, which removes the need for dedicated EMC hardware appliances. RP Virtual Edition is targeted at the storage admin and protects storage LUNs, is managed thru Unisphere (the storage-oriented interface), is deployed as virtual appliances on existing ESXi servers, and uses the embedded array splitter in VNX. It only supports EMC VNX today.
•RecoverPoint for VMs is targeted at the VMware admin or vAdmin and protects at the VM level, is fully managed thru vCenter (common interface for vAdmins), is deployed as a virtual appliance on existing ESXi servers, has an embedded I/O splitter within the vSphere kernel, and is storage agnostic and supports any SAN, vSAN, NAS or DAS storage arrays certified by VMware’s HCL. RP4VMs is a completely separate product from RecoverPoint – there is no upgrade, no downgrade and no interoperability with the existing RP products.
•RP4VMs leverages the field proven RP replication engine with over 20K units installed worldwide, 250 million run hours and over 1 EB of data protected and a world-class Global Services organization.
The ideal customer environment for RP4VMs is:
•Highly virtualized with VMware
•Want protection and recovery at VM-level granularity
•Centralized management thru vCenter
•Require zero or near zero RPO
•Better WAN optimization
•IT trusted and proven technology from EMC
RecoverPoint for Virtual Machines provides local and remote replication in combination with continuous data protection for per-VM recovery to any point in time. It supports both virtual disk types: VMDKs and RDMs. This diagram shows the architectural components and how RP4VMs works from the perspective of the primary user – the vAdmin.
The architectural components include: a VMware vCenter plug-in, a RecoverPoint write-splitter embedded in the vSphere hypervisor, and virtual appliances, all comprehensively integrated in a VMware ESXi server environment. RecoverPoint for VMs uses a journal-based implementation to store all the changes made to the protected VMs. Local protection provides a DVR-like rollback in time capability to any point, even to the last I/O transaction or just seconds before the data corruption occurred.
Unique capabilities of RP4VMs:
1. RP4VMs deploys a more elegant and true virtual appliance – a sealed box, delivered as a VM. It is optimized and tested under fixed pre-configured conditions and runs its own operating system. HA and distributed management is built right into the virtual appliance cluster architecture and the complexities are transparent to the user. Unlike our competition, who just runs a Windows app inside a VM that requires a running Windows OS, relies on the user to configure and properly install an HA cluster and requires a dedicated management server at each site.
2. More flexible deployment options: vRPAs can be deployed regardless of where the replicated VMs reside. There is no requirement to tie vRPA location together with VMs needing to be replicated. They can reside on any ESXi node and even can be deployed together on a dedicated ESXi node for hosting vRPAs. The flexibility in how you are able to deploy the vRPA provides differentiating features which is unmatched in the industry for VM replication solutions. Typically other solutions require that you couple replicating appliances to the workloads and are limited by the performance of their virtual replication VM.
3. Synchronous replication: EMC is the only vendor that offers remote protection over synch distances, from NYC to NJ. RP4VMs addresses the most demanding applications requiring zero RPO which none of our competitors provide today.
4. Superior WAN optimization for asynch replication – our built-in WAN optimization consists of deduplication, compression and advanced bandwidth reduction algorithms that reduce WAN network consumption up to 90%. WAN optimization also ensures replication robustness with an improved resiliency for jittery networks with up to 300 ms latency and up to 5% packet loss.
RP4VMs is designed to be customer installable and upgradeable and customers can go from download to VM protection in about one hour. RP4VMs uses the existing VMware ESXi server environment and infrastructure. We support vSphere and vCenter Server Release 5.1 U1 and 5.5.
Let me briefly talk about deployment. We will start with the splitters. The splitters are small pieces of code that work on the virtual SCSI layer of ESXi, they are deployed using a standard vSphere Installation Bundle or VIB, and are required to be installed on each ESXi node in an HA/DRS cluster where protected VMs may reside. It is recommended that the splitter be installed on ALL ESXi servers in the environment to assure all VMs can be protected regardless of vMotion activity.
Secondly, and the main component of the solution are the virtual RecoverPoint Appliances. These virtual appliances provide for efficient journaling and tracking mechanisms to enable everything from CDP-like capabilities and dataflow management of all replicated VMs. The vRPAs are deployed by installing an OVF file and then leveraging a Deployment Manager Wizard for final configuration and tailoring to the specific details of the environment.
With all of the splitters and vRPAs deployed, all that is left to do is deploy the management component of the solution. Management in RecoverPoint for VMs is done entirely from within VMware’s vCenter Web UI. Deployment is done by installing a VMware vCenter plug-in on each vCenter server that participates in the replicated environment.
vAdministrators are taking a more active role in protecting and recovering VMs housing important workloads, since they understand the business value of the application data and the required data protection service levels when the VMs are created. One of the pain points typically found with LUN level disaster recovery solutions is that management paradigms typically are not optimized for cloud environments and when managing these solutions, there is a lot of coordination and effort required between vAdmins and storage administrators to assure that VMs are properly protected at all times in an environment where VMs are moving constantly to adjust for workload demands. Not to mention all of this management usually occurs within separate storage or disaster recovery interfaces rather than integrated into the Cloud Management Framework. This not only causes day-to-day complexity in tracking VM location for assured protection, but also requires additional training by vAdmins who are critical in the disaster recovery process.
EMC’s RecoverPoint for VM’s management framework was designed to be fully integrated into VMware’s vSphere Web Client. There is no separate management interface to learn how to use.
Automation is critical when providing VM data protection services. As I just mentioned, VMs are being created, moved, and deleted every business day. Discovering and keeping track of which VMs are protected, is a hassle and can be prone to errors with manual procedures. RecoverPoint for VMs maintains a constant watch over the live configuration of VMs created and where they are hosted, and is able to protect a VM even when that VM and it’s workload are vMotioned to a different compute resource, or its virtual disk is Storage vMotioned to a different datastore. More than that, it automatically adapts to VM configuration changes so that when you add a VMDK to the source, a copy is automatically added to the replica and immediately starts getting replicated.
The vAdmin can create protection policy templates that define classes of service in advance – which encapsulate parameters such as RPO, protection window and WAN optimization. Then when provisioning the VM protection, the vAdmin can use the existing templates for easy provisioning, or define a custom policy for maximum flexibility.
RecoverPoint for VMs automates all of the internal protection virtual disk structures and journals that are required to properly protect VMs based on the protection policies set. There is no need to dive into RecoverPoint databases and manually provision journals.
Finally, RecoverPoint for VMs also provides for integration with the vCenter databases to properly create and configure replica VMs either on the local or remote site. There is no need to manually create VM entries at the DR site just waiting for a failover to occur. These can be automatically provisioned and created in vCenter on the fly. This capability is really designed with DPaaS in mind, enabling the vAdmin to properly configure both sides of the replication at once, even if the secondary site is in a public cloud at a CSP.
Along with being integrated within the native vCenter management interface, RecoverPoint for VMs management workflow is also incredibly easy to use. Wizard-driven workflows, allow for minimal error through guided VM protection configuration.
Here is an example of how easy and straight forward it is to apply a protection policy to a Virtual Machine leveraging RecoverPoint for VMs Protection Wizard:
First we specify if the VM should be added to an already existing Protection Consistency Group, <click> or do we want to create a new one with specific replication and protection SLAs.
Then we will specify if we want a new target VM copy to be created automatically, or if we want to select an already existing target virtual machine. RP4VMs will automatically create the replica and the journal as we just talked about. The source disks, destination disks, and the journals reside on virtual disks, which can be placed on any storage that is supported by VMware. Next, we also define the replication link policy and determine whether we want to have synchronous or asynchronus replication and what RPO we would like to define. The protection policies can be configured to meet specific protection, performance and cost goals and eliminate or minimize data loss. RP4VMs can dynamically switch between sync and async based on the available bandwidth.
For this walk through we will be creating a new target VM entry, so in this case, we simply specify which resource we would like that new target VM to be hosted from. This can be an individual ESXi Host, ESXi Cluster, or Defined Resource Pool in the environment.
Finally we will specify the datastore the target VM should be hosted on. And it is as easy as that. No complex configuration files, or storage interfaces to click through. Just a simple wizard and your VM is protected.
With RecoverPoint for VMs, the vAdmin is able to conduct automated disaster recovery orchestration tasks from within the vCenter plug-in. Automated DR workflows include:
•Recovery from logical corruption to any point
•Failover and failback of specific consistency groups or CG sets to any point
•Non-disruptive DR testing to any point
RecoverPoint for VMs allows the ultimate in flexibility for protecting applications by grouping VMs into Consistency Groups. Typically a consistency group represents a vApp, or a collection of VMs used by an application. RecoverPoint further offers an industry unique capability by supporting group sets: an additional layer of logical grouping, which provides consistency not only within an application but also across applications. For example, grouping a database server VM set along with a web front end server VM that leverages that database. These Consistency Groups and CG group sets are the same elements used for orchestration.
Taking orchestration flexibility further, the ability to manage Enhanced Group Sets allows you to perform Protection and Recovery operations at any logical grouping required. Test, failover and failback whole VM collections – from a single VM up to a whole site at once – at a single click of the mouse. And, many of the DR tasks are automated, such as storage mounts and restarts of VMs in the prioritized order to ensure recoverability of the applications.
For the ultimate flexibility, RecoverPoint for VMs uniquely allows these Consistency Groups and CG Group sets to span not only multiple ESXi nodes but also multiple ESXi Clusters, and even protect VMs from ESXi nodes being managed by different vCenters! It is this logical flexibility that allows our users to tailor the solution to their run book practices and custom applications.
RecoverPoint for VMs interoperates with VMware data migration and management services, such as vMotion and Storage vMotion, and is aware of changes to the underlying virtual infrastructure. Users can add or remove VMDKs from the protected VMs and RP4VMs will auto-provision VMDKs on the target side. The following changes can be made on the fly without impacting VM protection or the protection policies.
•Replicated VMs can be vMotioned
•Underlying storage can be Storage vMotioned
•Replicated VMs can be powered on/off or paused
With RecoverPoint for VMs you get simple, efficient and agile data protection for hybrid clouds.
First, enterprises and Service Providers can accelerate data protection as a service for private, public and hybrid clouds. RP4VMs enables IT to natively protect a software-defined data center. Through complete virtualization and VMware integration, vAdmins and app owners can now define protection policies as VMs are created, and these protection service levels are automatically enforced by the infrastructure that understands the business importance and resiliency requirements. Additional integrations into EMC’s Hybrid Cloud Solution and VMware’s vCD next year will complete the vision of delivering DPaaS for hybrid clouds.
Next, vAdmins are empowered with the visibility and control to manage replication using the VMware infrastructure they are accustomed to, with the object (VM-level) granularity and taxonomy they know, and via the management experience they already operate in.
Finally, enterprises can reduce protection cost and complexity by not creating another protection infrastructure, but using their existing ESXi environment to install and manage an agile data protection solution with built-in automated DR orchestration and superior WAN optimization to significantly decrease operational costs.
A demo I recorded showing RP4VMs protecting VM’s on XtremIO can be found below, I wanted to give an hugh credit to Idan Kentor from the DPAD that help me brining up the environment even though he was sick, you rock dude!