Photo of Douglas Jarrett

Amazon’s analysis of its extended cloud computing outage, as summarized in Richi Jennings IT Blogwatch, raises an important question for counsel advising clients negotiating agreements to procure cloud computing, content delivery and data communications services:  Should the agreement include a provision defining a service problem threshold and/or a series of problems threshold that triggers a right to terminate the agreement? We believe such a provision should be included, notwithstanding significant service provider pushback.

Several reasons underlie this position:

  • The concept of “cure” in standard breach and default provisions doesn’t really work.   Telecommunications and technology services agreements call for a service to be provided for a defined period of time.  When the service is unavailable or substandard during this period, the lost operating time cannot be recaptured.     
  • The widespread use of service level agreements (“SLAs”) impliedly recognizes that the “opportunity for cure” approach is largely irrelevant in circumstances when service is inadequate or unavailable.    
  • Standard SLAs are premised on the assumption that outages and periods of degraded service will be limited in duration and that credits and escalated response provide a sort of rough equity and reasonable compensation.
  • When an outage is extended, standard SLA remedies do not begin to compensate the customer for the disruption to its business.  

The potential termination of a problematic service relationship may give rise to the concern of “the cure being worse than the disease” because migrating from one services provider to another is often time-and resource-intensive and disruptive for the customer.  Thus, a service provider transition clause should always be included so the customer has adequate time to re-procure the service and migrate to a replacement provider.  During this transition, the service provider must be obligated to continue to provide and support the service and the customer must continue to pay for the service.

Clearly, reasonable contract provisions do not obviate the need for redundancy and contingency plans, as noted in a thoughtful blog post on the Amazon outage by Sharon Machlis.  On the other hand, a customer’s decision to procure back-up or alternative service should not negate a customer’s right to terminate an agreement when the primary service is degraded or unavailable for extended periods.  The service provider has put the Customer’s business at serious risk.  The customer purchased the insurance (redundant or back-up service) for its protection, not the services provider’s.