Most large disaster recovery outsourcers offer anything from a mirrored site with an RTO of greater than one hour to time-share sites with an RTO-RPO of fewer than 48 hours. For mirrored sites, there are severe distance constraints and DR site storage equipment is typically dedicated to one organization. For time-share sites, tapes or other media are shipped to the DR site and outsourcers supply servers, storage and networking to use. Some outsourcers also offer shell sites with only raised floor, power and cooling that can have an RTO-RPO less than one week. For shell sites, hardware must be shipped and installed prior to restoring applications and data.
What happens when a region wide disaster hits?
Timeshare sites are offered on a first come, first served basis. Many outsourcers provide alternate DR sites to be used when a designated DR site is full. However, any telecommunications and/or networking bandwidth services previously dedicated to support designated DR site business requirements must be re-routed to the alternate site(s) at the outsourcer's expense.
DR outsourcers should actively manage regional disaster risk, especially with respect to a 25 mile wide disaster centered around the nearest large city. Most outsourcers either limit this risk by not offering more services than can be simultaneously supported or oversubscribe and yet offer enough alternate sites to cover regional disasters. Oversubscription rates also often impact DR test scheduling availability.
Are the designated and alternate DR sites high availability data centers?
The Uptime Institute, the industry leading data center certification authority, defines four tiers of data center sites as follows:
- Tier I has basic infrastructure
- Tier II has redundant capacity components with single non-redundant distribution infrastructure
- Tier III has redundant capacity components and multiple independent distribution paths providing concurrently maintainable infrastructure
- Tier IV has fault tolerant infrastructure
Outsourcers should supply at least Tier III DR sites as defined by the Uptime Institute and should match or exceed the primary site availability level.
How much disaster recovery test time is supplied with your disaster recovery service?
Periodic tests or exercises are critical to any successful DR plan. Most contracts offer two days of test time per year but often more time can be purchased. Testing is critical in assuring the viability of the overall DR plan, both to the business as well as the outsourcer. Some service firms also offer detailed audits on DR exercises.
What hardware and software is supplied at the DR sites?
An outsourcer must supply hardware at DR sites compatible to hardware available at the primary data center. For instance, where data security is used at the primary site, compatible data encryption and key management services must be available at the DR site. Additionally, the outsourcer should supply software for all common operating systems and database servers. Special applications software is typically licensed to the primary data center owner and needs to be transferred to the DR site while running or testing there.
These five questions only constitute a starting point for an informed discussion with prospective disaster recovery services. More questions need to be asked regarding costs, DR and BC planning support, DR site personnel, etc. Disasters do happen – a DR plan using outsourced service providers may be a viable alternative.
About the author: Ray Lucchesi is president of Silverton Consulting, a storage, strategy and systems consulting services company.
This was first published in May 2010