Service Level Agreements (SLAs) crop up during most discussions of the cloud; and the conversation is typically not very conclusive. Some argue that you need an SLA so that you can use it to get back at your cloud provider when they fail. Others argue that they’re not worth the paper they’re written on. And a third group reckons they’re sort-of-useful in setting a benchmark for expectations of service. Cisco’s Rodrigo Flores had a nice piece on GigaOM this weekend which delves a little deeper. He argues that “the legalese in SLAs is useless,” but that “the vendor [should] articulate the service offering and expected performance clearly in a way that is visible, actionable by the customer and with clear metrics for recovery.” So ensuring the customer understands what service they can get is useful… weasel words about how little money they’re entitled to when your UPS fails less so. Today, most SLAs worry far more about the latter than the former, which is unfortunate. And then there are all the companies (some of them quite large) that offer no SLA at all. They argue that they should be judged on results rather than legalese, which may be a good point. But it makes it almost impossible for prospective customers to compare different offerings.