VMware - VMUG Montreal

Editor's Notes

• #6: Traditional business continuity solutions are complex and expensive.Business continuity can be implemented at the application level with application-specific clustering solution such as Oracle RAC, MS clustering for SQL, Exchange Database Access Groups (DAG), etc. App-level clustering usually provides best-in-class availability, but at a very high cost and complexity. There are also solutions that can be implemented at the infrastructure layer, typically backup and data replication. These tend to be more cost-efficient than app-level solutions, but can’t quite compare in terms of uptime and RTOs.
• #7: VMware provides a suite of Business Continuity solutions to offer holistic BCDR protection to all applications running on the vSphere platform. These solutions provide simple, cost-effective protection with a common solution for all your applications. The VMware BCDR solutions includes: Local availability products to protect applications against downtime of individual hosts. This includes vSphere HA and FT for unplanned downtime, as well as vMotion and Storage vMotion for planned downtime. Data protection solutions to back up entire VMs, including OS, application binaries, and application data, in a simple, non-disruptive manner. This includes the vSphere Data Recovery product, designed for smaller deployments, and the Storage APIs for Data Protection that enable third party backup vendors to integrate directly with vCenter and vSphere.vCenter Site Recovery Manager and the new vSphere Replication product to protect applications against site failures.The new Cloud Infrastructure Launch improves many of our BCDR capabilities, including vSphere HA, vSphere Data Recovery, vCenter Site Recovery Manager, and introduces the new vSphere Replication product.VMware vSphere™ is the leading virtualization platform that forms the foundation for building cloud infrastructures. It is specifically designed to holistically manage large collections of infrastructure – CPUs, storage, networking – as a seamless, flexible and dynamic resource pools that can be allocated to the right users and applications on-demand. vSphere comprises infrastructure services that transform server, storage and network hardware into a shared resource and application services that are built in and available to all applications that run on it.
• #8: VMware business continuity products provide the best of both worlds – similar RTOs and RPOs to app-level availability, but with the simplicity and cost-efficiency of shared availability services.VMware business continuity solutions provide protection at similar cost to traditional backup and replication – but with much better RTOs. For example, recovering an application from a traditional backup can take many hours or even days. Recovering an application from a VM backup only takes a few minutes, since the whole system – including OS, application binaries and data – are encapsulated as a single set of files. Similarly, doing a DR failover using traditional replication can take days, while recovering using SRM can take less than one hour.VMware business continuity provides similar RTOs to app-level availability, but at much lower cost and complexity. For example, recovering from a failed host with VMware HA takes a few minutes – similar to recovering from a host failure using MSCS. But VMware HA does not require dedicated standby systems, expensive software licenses, and complicated set up.h
• #9: We all understand that virtualization helps consolidate and reduce costs – and that is typically the first thing that comes to mind when thinking about virtualization. But it’s also interesting to note that for our existing customers, the #1 objective for virtualization is consistently to improve BCDR.VMware vSphere is inherently a better platform for BCDR that offers many new BCDR options. And customers that know VMware well make it a priority to leverage these inherent capabilities to improve their BCDR configurations.
• #10: Let’s do a quick overview of SRM for disaster recovery
• #11: Key Points:Disaster recovery in the traditional physical x86 world is hard for the following reasons:Easiest and most reliable recovery requires complex server and recovery site infrastructureThe recovery process is also typically fairly complex due to hardware dependencies, application dependencies, etc.Due to the complexity of the process, having accurate and up-to-date documentation (a disaster recovery “runbook”) is essential, as are ensuring that everyone is thoroughly trained on recovery procedures and executes them correctly when neededAs a result, recovery requirements often fail to be met: recovery takes too long, is often unsuccessful, and requires significant investment of IT $$ and staffScript:Traditional disaster recovery plans depend on a very complex set of processes and infrastructure: duplicate datacenters, duplicate server infrastructure, processes for getting data to a recovery site, processes for restarting servers, processes for reinstalling operating systems, and so on. Because disaster recovery can be some complex, organizations often find themselves unable to provide good protection to more than a privileged few of their production workloads, leaving other workloads (e.g. file/print servers, internal web servers, departmental applications) unprotected or poorly protected.Because of the complexity of disaster recovery plans and infrastructure, organizations are heavily dependent on significant amounts of personnel training, on the accuracy and completeness of thick paper “runbooks” that document the recovery process, and on perfect execution of the recovery process when an outage does occur. Because testing is disruptive and expensive, organizations have a limited ability to ensure that all of their training, documentation, and execution is practiced and can successful recover their IT services.As a result of these challenges, tests of recovery plans often fail; basic recovery of critical workloads – if successful at all – often takes days or weeks; and a significant amount of IT time and resources are consumed by managing and maintaining recovery plans. In short, most firms fail to meet the continuity requirements set by their organizations.
• #12: I want to review 3 key features of virtualization that lead to better disaster recovery.Consolidation. Server consolidation means doing more with less. You’re reducing your physical footprint, which also means you can streamline your disaster recovery plans and standardize your recovery process. Hardware Independence. This means being able to recover onto any x86 hardware. So, you have the flexibility to buy different servers for your recovery site, or even fewer servers. Continuing to virtualize your production site will free up additional machines that can then be moved over to your recovery site. Encapsulation. Because virtualization captures everything about a server into just a few files on disk, the real benefit here is mobility. The ability to move your VMs wherever you want. You can back them up in the same way you currently protect your other files. You can also replicate them to your disaster recovery site, so that they will be available when you need to recover from an outage.
• #18: SRM provides very broad application coverage.The RPO is set by the replication product in use. Since SRM supports a very broad range of replication products, the RPO can be set anywhere from days to synchronous replication, depending on business needs, replication capabilities, and network bandwidth.The RTO will depend on the actual configuration, but is typically in the range of 30 mins to a few hours. Factors that will influence this include the number of VMs that need to be recovered, the number of hosts available to recover those VMs, and the replication technology.SRM can be used to protect the vast majority of applications. The only applications that can not be protected with SRM are applications that need better RTOs than 30 minutes to a few hours. Applications that need continuous or near-continuous RTOs will need to be protected using geo-clustering.
• #19: Key Points: Site Recovery Manager can be used in a number of different failover scenarios In particular, SRM helps you to make better use of your investment in a second site—you can use that second site for other workloads when you aren’t in a disaster recovery scenario rather than just having it sit idleScript:Site Recovery Manager can be used in a number of different failover scenarios:Active-Passive: Site Recovery Manager absolutely supports the traditional active-passive DR scenario, where a production site running applications is recovered at a second site that is idle until failover is required. Although the most common configuration, this scenario also means that you are paying a lot of money for a DR site that is idle most of the time.Active-Active: To make better use of the recovery site, Site Recovery Manager also enables you to leverage your recovery site for other workloads when you aren’t using it for DR. Site Recovery Manager can be configured to automatically shutdown or suspend VMs at the recovery site as part of the failover process so that you can easily free up compute capacity for the workloads being recovered.Bidirectional: Site Recovery Manager can also provide bidirectional failover protection so that you can run active production workloads at both sites and failover to the other site in either direction. The spare capacity at the other site will be used to run the VMs that are failed over.Local Failover: Although less common, some of our customers need to be able to failover within a given “site” or campus, for example when a storage array failure occurs or when building maintenance forces you to move workloads to a different campus building. These customers are leveraging Site Recovery Manager to perform these failovers.
• #22: Today, DR coverage is often limited to Tier 1 apps in larger datacenters. In many cases, Tier 2 / 3 apps ad smaller sites do not have true DR protection, but are only protected with backups. This is because traditional DR protection is cost prohibitive and too complex to be broadly applied to all applications and sites.Unfortunately, this level of DR protection leaves a substantial business risk as day-to-day activities still rely extensively on the T2 / 3 apps and smaller sites. Ideally, organizations should have a simple, cost-effective, reliable DR plan in place for all their applications and sites.
• #24: vSphere Replication has been designed specifically to provide very cost-efficient, simple, yet powerful replication for SRM deployments.It is more cost-efficient because it reduces both storage costs and replication costs. At the storage layer, vSphere Replication eliminates the need to have higher end storage arrays at both sites. Customers can use lower end, different storage across sites, including Direct Attached Storage. For example, one popular option is to have Tier 1 storage at the production site, but lower end storage at the failover site, such as older arrays or less expensive arrays. vSphere Replication is also bundled with SRM at no additional cost, eliminating the additional cost of storage-based replication licenses.vSphere Replication is also inherently simpler than storage-based replication. Replication is managed directly from vCenter, eliminating dependencies on storage teams. And it is managed at the level of individual VMs, making the setup of SRM much simpler.Despite its simplicity and cost-efficiency, vSphere Replication is still a robust, powerful replication solution. It can provide 15 minute RPOs, and gives the flexibility to set RPOs between 15 minutes to 24 hours. It tracks changed disk areas and only replicates the latest deltas to increase network efficiency, and will scale upwards of 500 virtual machines. It does however have a few limitations, such as the lack of support for automated failback (in this initial release), file-level consistency, no support for FT, linked clones, templates, and physical-mode RDMs.
• #25: vSphere Replication is a great way to expand DR protection to Tier 2 / 3 apps and smaller sites.Storage-based DR protection is fairly expensive, with the cost mostly driven by storage capacity on Tier 1 storage arrays and additional replication licenses. The typical storage, replication and SRM cost is in the range of $2,000 per VM. While this is much better than physical DR, it is still fairly high and can be cost-prohibitive for less business-critical environments.vSphere Replication is much more cost-effective. By supporting the use of lower-end storage arrays, eliminating the need for dedicated replication licenses, and offering lower-cost SRM Standard licenses, the cost per VM is reduced by 3X to approximately $600 / VM.The much lower cost per VM enables organizations to expand their DR protection to many more applications and sites.
• #30: Setting up an SRM recovery plan is done in 5 simple steps.1 – The user maps resources at the production site to resources at the failover site. This includes resource pools, vSwitches, and VM folders. The VMs will automatically be mapped to the appropriate resources upon failover.2 – The user selects which virtual machines and virtual machine protection groups to include in the plan. This enables users to set up plans for failing over only part of the protected VMs.3 – The user selects which low-priority VMs to suspend at the recovery site – for example test and dev VMs that will be suspended to make room for the recovered VMs.4 – The user specifies the boot sequence of recovered VMs – for example, first the database, then the application tier, and finally the web tier.5 – Finally, the user can customize re-IPing of VMs to comply with the network at the failover site.Optionally, users can add messages, pauses for execution of manual steps, and custom scripts for example to update DNS servers with new IP addresses.SRM eliminates many of the traditional recovery steps by automatically coordinating recovery across infrastructure layers. For example:Individual hosts no longer need to be re-configured at the recovery site Users no longer need to worry about recovering entire systems including the OS, application binaries, and application data – since the entire system is recovered with the VM Coordinating of replication and storage is done automatically. Users no longer need to worry about stopping replication, presenting the LUNs to vSphere, and registering the VMs into vCenter There is no need to reconfigure physical switching infrastructure.
• #31: Customers often wonder whether SRM enables application consistent recoveries. In other words, whether the application can be recovered in a clean state, or whether the application is recovered in a crash-consistent state.The short answer is that it is possible to ensure application-consistent recoveries. The application consistency is enabled by the underlying replication technology. Many replication vendors have management products for this purpose – such as EMC’s Replication Manager or NetApp’sSnapManager.Typically, these products place an agent in the VM that will quiesce the VM prior to executing the replication. The quiesced VM at the recovery site is then used for recovery purposes.In this initial release, vSphere Replication will only provide file-level consistency, and not application consistency.Even when application consistency is not ensured by the replication product, the ‘planned migration’ workflow, new with SRM 5, ensures application-consistent migrations for planned migrations. This is the case for both storage-based and vSphere Replication.
• #42: SRM 5 now comes in two editions: Standard and Enterprise. Standard is designed for smaller environments, and can be used to protect up to 75 VMs (per site and per SRM instance). Enterprise is designed for larger environments with more than 75 VMs to protect. Both editions are full-featured and include vSphere Replication, automated failback, and planned migration.The tiered SRM editions provide a particularly attractive price-point for SMBs and Remote Offices – at $195 per VM. This is a steep reduction from the standard SRM 4 list price of $450 / VM. This more attractive price-point will enable smaller organizations and sites to leverage vSphere Replication and significantly reduce the overall DR costs.SRM Enterprise is priced at $495 / VM, a very small increment to SRM 4 which now includes vSphere Replication and the other new SRM 5 capabilities.
• #43: VMware provides a comprehensive set of BCDR services.The VMware SRM Jumpstart helps organizations get started with SRM. It’s a short 3-day service offering, designed to do an initial PoC / on-site installation of SRM, integrate SRM with the storage arrays, and set up a simple recovery plan.VMware also offers a custom BCDR Plan and Design service, intended to provide a holistic BCDR service covering local availability, data protection, and disaster recovery.