Summary at the top = good session, nothing earth-shattering but nice development of current DR ideas….like the direction VMware is going here and matches up with where Varrow is going as a company. Hosting.com also had a very impressive screenshot demo of a per-VM, 2-click DR service.

This is all futures and nothing committed here, etc. etc.

• Standard SRM Overview — replace manual runbooks, allow non-disruptive business-hours testing, etc.
• Today = array based replication. 28 SRA’s (array adapters from storage vendors)….impressive.
• Future 1 = evolve SRM from being VM-centric to service-level centric and application aware.
  • Protect “Exchange” at application provisioning, specify service level, and you’re done.
  • I like this…but this isn’t dreaming very big yet….
• Future 2 = direct public cloud integration more easily.
• SRM 5 Features
  • Automated Failback — single button failback all recovered VMs, interfaces with storage to automatically reverse replication.
  • vSphere Replication — native ESXi host-based replication at per-VM granularity, storage agnostic, replication between heterogeneous storage arrays.
  • Planned Migration — graceful shutdown of protected VMs with storage sync prior to failover, resulting in application consistent VMs on recovery.
  • New User Interface — single pane of glass across DR sites, redesigned workflows, IP Customization UI (finally!), shadow VM icons, history reports who initiated DR action.
    • No more two interfaces..made easier by vCenter Linked Mode but still 2 interfaces.
    • Much more granularity in reporting and tracking.
  • Recovery Enhancements — Priority Groups & Inter-VM dependencies, In guest scripts.
    • Went from 3 to 5 priority groups, within a single priority group can also force a VM to depend on another (very nice…lots of flexibility here).
  • Performance, Scalability, and API Improvements
• 2012 and Beyond — Cloud DR, Policy Based DR — automating DR. vMotion Anywhere Support, SRM as Migration Manager, RTO Reduction (even more!), Desktop DR, Per VM Protection with SRAs.
• Cloud DR — so what is it? Goals are….
  • Goal = enable same level of DR protection as currently enable with SRM but for any workloads running primarily in the cloud….vCloud Director based environments.
  • Provide simplified, self-service DR solution to existing single-site vSphere and vCloud Director customers.
    • “I can’t afford a DR site….give me something cheap or easy or fast or….”.
  • Reduce DR complexity and go even further towards a “single click” DR solution.
• Who does cloud DR apply to? Well, anybody of course….but….
  • Single-site vSphere customers — no replicated storage, also SMB customers.
  • Public Cloud providers — can you smell the business opportunity here? I can…
  • Private Cloud customers — the big guys….you’re the 500+ VM people and you know who you are (maybe 200+ VMs).
• Cloud DR Phases — deploy it (well, buy it first), configure, Initiate Testing, Continuing Monitoring of DR Operations.
• Use Cases
  • Single site vSphere 
    1. Purchase DRaaS subscription from vCloud Service Provider (VSP)
    2. Org VDC (virtual datacenter) configured in vCloud as DR container
    3. Obtains DRaas site activation/configuration details from VSP — this consists of a huge config file (likely xml) that automates all the DRaaS setup work.
    4. DR software installed at vSphere site and auto-configured by importing config file.
    5. VMs added to DR Ptorection Group(s) and assigned to subscribed Recovery Plans using vSphere client.
       • Canned Protection Levels and you drag’n drop to choose what level of protection you want per VM.
  • DR *of* the Cloud — larger private cloud (aka vCloud Director) at the customer site
    1. Provider Admin deploys cloud DR infrastructure at both vCloud sites.
    2. Org admin selects DR protection as a catalog service (can you say ITIL? 😉 during VDC creation and assign DR levels.
    3. More steps here….lost all this.
• Benefits
  • Lower cost per VM for DR protection, CapEx savings as don’t need to stand up duplicate infrastructure, OpEx savings around day to day DR site operations
    • All that allows you to protect additional workloads.
  • Lower RTO for DR protecting workloads — compared to the DR options for single-site environments that is.
  • Easier easier easier….say it with me now.
• Key Cloud DR Features — this is a laundry list but we need all these things…
  • VM-level protection granularity, Multi-Tenancy, Self-Serviceability, Storage Agnostic, VM Portability, Roll-based Management, Scalability, Extensibility, Simplified Deployment and Management, Security (vmdk encryption), RAS (Reliability, Availability, Serviceability)

And now a big ad for hosting.com!! Seriously though, I appreciate seeing something built out around all the theory stuff above. Craig the CTO from hosting.com talking (who used to work for Sungard…..the ultimate automated DR nemesis, right?) He’s simplifying his deck to stay in the time frame….3 myths to start out.

1. Myth #1 – Most Disasters Are Caused by Catastrophic Events
   • Far more problems caused by Clinical Disasters rather than Catastrophic Disasters.
   • No patience for Clinical Disasters….which are Power Failure, IT hardware failure, Network failure, IT software failure, etc.
   • 73% of disasters are Clinical Disasters.
2. Myth #2 – Direct Costs constitute most outage costs.
   • Indirect costs are more than 2x the direct costs.
   • Compliance is huge here — auditors won’t tolerate being out of compliance while in DR state and could get big fine.
   • Indirect costs = reputation, future sales, etc.
3. Myth #3 – All Disaster Recovery Plans are Created Equal
   • Only 1/3 of applications are deemed mission critical….so why have one DR plan?

Hosting.com focuses on High Availability, Recovery Capabilities, Speed of Provisioning, Cost Alignment…..you fill in the definitions.

Screenshot demo walkthrough of hosting.com DR service…see the PowerPoint for screenshots.

• Unique customer dashboard after login — tab called “Recovery Services”.
• Select VM, order protection, choose site to replicate to.
• Look at a VM, click “Add Recovery” button, choose the DR site, done.
  • This is very, very impressive…..one-click DR on a per VM basis.
• Compliance-driven log options.