Cloud Computing is a paradigm already established to support the deployment of applications and services without the need to rely on costly on premise infrastructures. A unique service provider sometimes owns the data centers supporting the cloud, but there is a trend to a fully distributed architecture, where several data center providers will have to cooperate through the interconnection of their computation and storage infrastructure. The interconnection network is a fundamental part in the performance of such a distributed cloud environment. As different services require different computing and storage capabilities, they also require different performance parameters related to the network, like bandwidth, delay or jitter. Allocating resources from different data centers and their interconnecting network systems to satisfy the needs of heterogeneous service requests and all in an automated and efficient manner is a challenge that current cloud computing platforms are not able to fulfil.
In this project, we consider that computation, storage and networking resources have to be considered as a whole to be allocated to service requests. Each service request has to be satisfied irrespectively of other requests arrived in the past and those that can come in the future. Handling each service as something that is independent of others, but at the same time sharing the same substrate infrastructure, drives us to the concept of virtualization. Virtualization is assumed to be applied to target resources of any type, so that allowing to allocate e-virtual resources to heterogeneous service requests. The set of virtual resources that will enable a service to be deployed fulfilling its specific requirements over time, as it were the only existing one in a given computation and networking infrastructure, gives rise to the Lightweight Slice Defined Cloud (LSDC) concept coined in NECOS. The innovation of NECOS will be to materialize the LSCD concept by means of a platform exhibiting the following four distinguishing characteristics:
It presents a new service model – the Slice as a Service. A slice is beneficial as the management software can dynamically map service components to a slice. A slice is a grouping of physical or virtual (network, compute, storage) resources that can act as a sub-cloud, sub-network and can accommodate service components, seemingly independent of other slices. The enhanced management for such a slice enabled infrastructure creates slices on-demand, and (re)configures them as appropriate to provide the end-to-end service. The slice management takes over the control of all the service components, virtualized network functions and system programmability functions assigned to the slice.
It enables the configuration of slices across the physical resources in the cloud networking infrastructure, in order to better accommodate the various service demands. Adaptions and reconfigurations, including interoperability and data portability issues, are done at a per slice level, rather than for a whole cloud. This allows a much better adaptation to the requirements of the service requests. This configuration is achieved through using specially designed software that can describe and manage the various aspects that comprise slices within the cloud environment.
It allows each aspect that comprises the cloud environment – from the networking between virtual machines, to the SLAs of the hosted applications – to be managed via software. This reduces the complexity related to configuring and operating the infrastructure, which in turn eases the management of the cloud infrastructure. Such infrastructure tends to be very large scale, commonly composed of thousands of servers and network elements, supporting tens of thousands of virtual machines, virtual networks, and applications.
It utilises lightweight and uniform management and virtualisation systems, with small footprint components, deployable on a large number of small servers and cloud systems both at the core and the edges of the network. These lightweight elements enable the integration of core data centers and mobile edge into cloud networks ensuring data portability and compatibility between clouds.
The LSDC approach will be manifested through the following results:
Information models for slice specification
Information models and methods needed to specify a slice in which the required cloud resources and services are described, discovered and consumed. In order to use these information models, users will leverage an integral part of this exploitable artefact. This project outcome can be of interest to system designers and manufacturers of slice-based products.
Lightweight, flexible and real-time monitoring facilities for cloud federations
Monitoring of physical and virtual resources is necessary for platforms that aim to optimize the use of resources. In that sense, this component can be of interest in any management system of virtual resources. System designers and manufacturers will be the most natural markets to target.
Real-time orchestration of the slice lifecycle
This component constitutes the core of the LSCD platform and it is in charge of deciding where, how, and when to place slice-level elements (i.e., virtual machines and storage) over the federated cloud. It will be based on algorithms of machine learning, optimization techniques, and feedback-based loops.
This module will concentrate an important part of the effort of the project, including the prospect to create new techniques for slice orchestration across the cloud federation. For that reason, we understand that this module can be easily fragmented into their constituting algorithms and each one of them can be of interest to the scientific community. On the other hand, system designers and manufacturers can be interested in the module as a whole.
The LSCD platform
This is the main objective of the NECOS project. It will accept user’s requests for service deployment and it will trigger the process of allocation of the service to an existing slice or the creation of a new slice in case that it is necessary. The LSCD platform will also contact with other counterparts belonging to other administrative domains.
The LSCD platform is aimed to cloud service providers and network providers that want to cooperate to provision services tailored to diverse user-specific needs.