Cloud computing potentially ushers in a new era of computer music performance with exceptionally large computer music instruments consisting of 10s to 100s of virtual machines which we propose to call a ‘cloud-orchestra’. Cloud computing allows for the rapid provisioning of resources, but to deploy such a complicated and interconnected network of software synthesizers in the cloud requires a lot of manual work, system administration knowledge, and developer/operator skills. This is a barrier to computer musicians whose goal is to produce and perform music, and not to administer 100s of computers. This work discusses the issues facing cloud-orchestra deployment and offers an abstract solution and a concrete implementation. The abstract solution is to generate cloud-orchestra deployment plans by allowing computer musicians to model their network of synthesizers and to describe their resources. A model optimizer will compute near-optimal deployment plans to synchronize, deploy, and orchestrate the start-up of a complex network of synthesizers deployed to many computers. This model driven development approach frees computer musicians from much of the hassle of deployment and allocation. Computer musicians can focus on the configuration of musical components and leave the resource allocation up to the modelling software to optimize.