HEAT Wrapper

Heat Wrapper implements the interface of both a compute and storage wrapper using the OpenStack cloud manager to offer access to computational and storage resources. In the current implementation of the service deployment flow, the FLM take care of the deployment of each VNF separately, and so each VNFD is sent separately to the IA. Therefore the Heat Wrapper must be able to carry out the deployment of each function independently from the others. The translation between TANGO Network Service (NS) and VNFs models to Heat template has been designed to allow this VNF-per-VNF deployment. In fact, the IA leaves control for the deployment and lifecycle management of VNFs to the MANO framework, also allowing the VNFs of a NS to be deployed in multiple NFVI-PoPs. In order to do this, the translation process of a VNFD must be independent of other VNFDs and from the NSD. Usually, OpenStack Heat stacks are used to represent a complete service composed of several virtual machines and networks. In our model, a stack running on an OpenStack VIM represents the subset of VNFs of a NS which have been placed on that specific VIM. Moreover, since a NS virtual link can span more than one PoP, it cannot be directly mapped 1-to-1 to a Neutron Virtual Router, so intra-VNF connectivity is completely delegated to Networking VIM Wrapper, leaving to the Compute VIM Wrapper in general, and OpenStack Heat in particular, the task to create basic layer 2 and layer 3 connectivity where SFC or other network policies could be enforced later on by a Networking VIM Wrapper.