International telecommunication union
Distributed cloud for service providers
Download 0.91 Mb.
|
- Navigate this page:
- 9.2 Key characteristics
9 Distributed cloud for service providers9.1 IntroductionCloud computing services are typically delivered via datacenter-like infrastructure. There are numerous types of datacenters with their respective advantages and disadvantages. The Figure 6 depicts different types of datacenter: 
Figure 6 Type of datacenter Various 5G applications have different requirements in terms of latency, throughput, and cost. Latency sensitive applications such as gaming, voice, video, and telepresence can benefit from the characteristics offered by regional, metropolitan, or access clouds. In a similar fashion, best-effort none-latency sensitive applications can profit from the cost advantage associated with national or continent-wide datacenter solution. We can expect that in 5G some operators will leverage the distributed nature of their communication network by highlighting the benefits provided by regional or metropolitan clouds.19 A distributed cloud infrastructure supports the deployment and migration of applications between any of these types of cloud infrastructure. The capability to deploy the same application in either a national datacenter or near the network edge can bring unique advantages to operators. As example, a stock trading application would certainly gain from much shorter latency provided by a metropolitan-area rather than a centralized national cloud infrastructure. A telecom cloud system management in 5G logically sits above the various types of datacenters. Such cloud manager decides in which particular site and processor pool an application would be initiated to fulfill the characteristics and cost defined in the Service Level Agreement associated with it. A distributed cloud concept which encompasses processing and network resources is fundamental to the success of IMT-2020 network.
Figure 7 distributed cloud The key characteristics of distributed cloud are:
A consumer can unilaterally provision capabilities, such as server time, network storage and network bandwidth, as needed without requiring human interaction with each service’s provider.
Capabilities are available over the network and accessed through standard mechanisms that promote use by heterogeneous thin or thick client platforms (e.g., mobile phones, laptops, and PDAs).
The provider’s computing resources are pooled to serve all consumers using a multi-tenant model, with different physical and virtual resources dynamically assigned and reassigned according to consumer demand. The customer generally has no control or knowledge over the exact location of the provided resources. Examples of resources include storage, processing, memory, network bandwidth, and virtual machines.
Capabilities can be rapidly and elastically provisioned to quickly scale up and rapidly released to quickly scale down. To the consumer, the capabilities available for rent often appear to be infinite and can be purchased in any quantity at any time.
Resource usage can be monitored, controlled, and reported providing transparency for both the provider and user. In addition to these characteristics, the distributed cloud provides higher resilience and availability, greater security and the support of differentiating service classes, which will drive the deployment in order to guarantee the described Operational Level. Gap analysis The existing IMT network has its limitation and lacks of flexibility and agility for deploying the network functions and applications at any location where the performance and user experience would be optimized. In order to meet the extremely various demands of the services in 5G, for example, from ultra-low latency to high-latency tolerable service, distributed cloud technology provides a viable solution. To realize its benefits, the following gaps would need to be filled: (1) Distributed Storage Services that provide uniform, system-wide, distributed block storage for the applications (e.g., OpenStack’s Swift and Cinder subsystems) (2) Networking Services that SDN enables such as cloud-wide, virtualized connectivity, both at L2 and L3 levels, such as OpenStack’s Quantum, (3) Distributed Compute Services that manage VM’s, doings tasks such as start, stop, migration and supervisions of VMs is to be performed by a cloud computing service (e.g., CloudStack and OpenStack’s Nova) and (4) Cloud management API for applications on top of the cloud infrastructure (e.g., OpenStack) for application deployment, migration and portability. Although it is ideal to have all applications running inside VM’s, reality, at least in the short term, dictates that some tasks must continue to execute on non-virtualized or specialized hardware. In order to limit the extra OPEX burden such system anomalies represent, it is still necessary to provide these environments with an API that makes it possible to manage them the same way (i.e. by abstracting and presenting a uniform interface to the applications) as is done with VM’s, so that it is still possible to keep parts of the management APIs (loading, start, stop etc.) uniform and identical to the ones as in VMs.
Directory: ITU-T -> focusgroups focusgroups -> International Telecommunication Union focusgroups -> Innovation of ict in Developing Countries focusgroups -> International Telecommunication Union focusgroups -> Telecommunication focusgroups -> Focus Group on Smart Grid focusgroups -> Telecommunication focusgroups -> Telecommunication focusgroups -> Telecommunication focusgroups -> Innovation of ict in Developing Countries focusgroups -> Innovation of ict in Developing Countries Download 0.91 Mb. Share with your friends:
|
