Whether the word 'disaster' causes one to visualise a bomb, flood or power failure, a common theme is that all organisations now accept that they must be prepared for the worst and have plans established in order to preserve business continuity. Pasteur is quoted as saying that chance favours the prepared mind; in the world of disaster recovery planning, this rings very true. Business continuity plans range in their complexity and coverage, and this makes it very important to define what is within and without scope of a disaster when one looks at the invocation criteria for a disaster recovery plan. For example, during the years leading up to the year 2000, there was many an argument as to whether a year 2000 failure was within scope. The disaster recovery service companies at the time argued strongly that Y2K rectification services were outside of standard disaster services (in order to escape the technical and financial burden of dealing with the Y2K issue).
Disaster recovery has an interesting relation with the concept of force majeure (the "acts of God" clause tucked away in the boilerplate of a standard contract whereby parties are excused their obligations for the duration of a force majeure event). When it comes to disaster recovery, it can be precisely those acts of God which you will want your contract to cover and to provide disaster recovery services for. The normal contractual rules are therefore subtly different in disaster recovery contracts.
Often disaster recovery services are outsourced, i.e. an organisation needing disaster services will go to a specialist for a provision of such services. Why outsource? Well it's the usual reasons such as to achieve costs savings (running a parallel site just for disaster recovery is rarely economical for a business), find a better service, for reasons of corporate policy and to overcome recruitment issues. The decision whether to outsource (whether outsourcing disaster recovery or outsourcing ICT services generally) is made following a cost-benefit analysis of such issues.
Contractually, disaster recovery contracts tend to be IP-rich contracts in that the intellectual property elements are usually key to the service provider's meaningful ability to provide disaster recovery services. In the normal way, an organisation's ICT will be a mix of numerous items of hardware and software, often under lease or licence arrangements.
One matter that is unlikely to be permitted by standard package software licence agreements is third party use: whilst you may be able to use the software licensed to you, the licence is unlikely to allow you to give it to someone else to help you use it. However, for a disaster recovery situation this may be exactly what you need to do.
Additionally, a number of software companies restrict their licensing on a site-by-site, premises-by-premises basis. It may therefore be necessary to go back to the software company to obtain permission to use the software or to replicate the software offsite, as well as to allow third parties to use it.
So what are the key issues in a disaster recovery contract? The main issues of concern tend to be:
Timescales for performance;
Acceptance of deliverables;
Charges and payment profile;
Invocation criteria; and
Scope of facilities.
Timing is very important to keep tight the safety net provided by DR facilities. When one is setting up a mirror installation at a disaster recovery site, for example, the work to set up such installation will need to be done speedily. As such, the contract will need to have a contractually enforceable implementation plan backed up by legal redress (such as liquidated damages) to compensate for delay. Additionally, you may wish to make the timescales "of the essence" to make termination easier should timescales be missed. Such a provision would be appropriate in a time-critical contract such as this.
Acceptance and the ability to reject is very important and often has a direct relationship with the payment profile. Typically an organisation will not wish to pay for disaster recovery services (or the installation work that has been undertaken to establish those services) until the installation has proven itself through an acceptance testing process (which might in disaster recovery cases include simulated disaster events). Acceptance is generally the point at which you lose your right to reject and should not be taken lightly. As such testing should be performed jointly by customer and supplier (to avoid disputed results).
Customer organisations should take care to avoid acceptance being deemed to have occurred, whether by 'live' use of the installation or otherwise. Look out for this in DR contracts and insist that they allow for upon positive affirmation or positive rejection - in both cases in writing.
It is often suitable in a disaster recovery contract (as with many other IT contracts) to stage payments to provide a monetary incentive to complete the required services. Often in DR contracts there will be a separate fee for invocation of the disaster services in addition to payment for set up.
The trigger events for disaster recovery need to be set out clearly. For example, your level of disaster recovery service may kick-in for system failure continuing for more than 4 hours, or only following the occurrence of a pre-defined event. The process for who contacts whom as well as the practicalities of how, for example, backup tapes become live tapes need to be set out clearly.
Scope of Facilities
Disaster recovery contracts tend to offer either dedicated facilities or shared facilities. Dedicated facilities, which are available exclusively to the customer, will cost more than a share of general facilities. However, by securing dedicated facilities an organisation will have purchased a guarantee of space and the ability to continue its organisation seamlessly. By contrast, if the facilities are shared there is usually a 'first come, first served' rule. If the disaster is widespread affecting a number of organisations all with a share in the facilities, an organisation which invokes its right to the services late may be left out in the cold.
(a) Term and Structure:
A typical disaster recovery outsource contract might be for a 5 year term, but should, practically, allow a right of extension. There will be initial service provision to set up the disaster recovery installation, with acceptance tests.
Many organisations will secure a split of both shared and dedicated services. An organisation should ideally seek "most favoured customer" status such that it gets first call on any shared services upon invocation. The customer should always ensure that the disaster recovery company informs the customer when parts of the shared facilities are already in use.
(c) Ongoing Service Maintenance:
There should be regular tests of the disaster recovery facilities and clear rights to claim consequential loss if the disaster recovery services fail. Finally, a useful mechanism for contracts of this type is an upgrade fund. Here, the customer agrees to deposit monies periodically into a separate account which are applied to the purchase of hardware and software upgrades in order that the disaster recovery installation be kept up to date and commensurate with the main installation.
A version of this article was published in the April 2002 edition of MIS magazine.