Is this for an apples to apples bid specification? If it is, unless you can ensure all bidders provide the same level of solution based on very specific requirements (which is typically not the case), I would hard specify the type of solutions that are acceptable then make it my business to learn everything and anything associated with those solutions. Otherwise you could potentially open yourself up to a bidder that doesn't understand the level required and either under design accidentally or intentionally to win the project.
As for general requirements, this truly depends on the requirements of the client. What if a failure occurs and the tape back up is off site? Can the client wait 24 hours for tape retrieval Or is it needed immediately? If high availability clustering and the disaster is in the server room (fire/water) then likely the backup will be destroyed as well. Hybrid cloud system begs the question of sensitivity of the data being stored and availability to send the data to the cloud (T1 won't likely work for numerous multimega pixel cameras).
I would identify the acceptable (hard) perameters and keep the requirements clear for the bidders so you can control the outcome for the client.
Thanks for the response – I am looking for information on the planning side rather than the solutions side of a DRP. Consultants – when developing specifications for surveillance DR, can you give me the highlight procedures / general requirements of a plan? For example – since there are various methods to achieve data recovery (high availability clustering, on/off-site tape backups, hybrid cloud systems, distributed data replication, etc…) how do you begin building a specification? Is it generally based on the company’s business continuity plan? What if they don’t have a continuity plan?
Is it typical to leave it up to the integrator to specify the metrics used in a disaster recovery plan?
We have used Pivot3's VStac where data is not only striped across drives, but also across servers. They call it RAID6x, where a group of servers can withstand an entire unit failure and several drive failures simultaneously, but it requires a minimum of 3 vStac units.