新型云网融合编排与调度系统架构与分析

Architecture and analysis of a new cloud-network convergence orchestrator and scheduler system

算力时代,新型算力业务的异构数据处理需要云边协同的泛在化算力支持,同时算力也在向包括算力、存力、运力与能力的新型算力演进,以满足新型算力业务对算力服务的多样化需求,这就需要云网融合编排调度系统具备自动化、智能化、孪生化的新型算力编排调度能力,实现云网全局拓扑下,业务目标服务等级协议与云网资源组合的联合最优解,从而提升算力效能。在分析现有云网融合编排调度系统不足的基础上,提出了一种自动化、智能化与孪生化的新型云网融合编排调度系统架构,给出了典型场景,并提出运营商云网融合编排与调度系统的演进路径建议。

In the era of computing power, heterogeneous data processing of new computing power services requires the ubiquitous computing power supported by cloud edge collaboration. Meanwhile, computing power is evolving to the new computing power including computing power, storage power, transportation power and capability, so as to meet the diversified demands of new computing power services. This requires the cloud-network convergence orchestrator and scheduler system to be equipped with automatic, intelligent and digital twinization capabilities of new computing power, so as to figure out the quantitative relationship between target SLA(Service Level Agreement) and resource portfolio of cloud and network conditionally dependent on the global topology of cloud-network convergence, in order to improve computing efficiency. On the basis of analyzing the deficiencies of the existing cloud-network convergence orchestrator and scheduler system, proposes a new cloud network convergence orchestrator and scheduler system architecture which is automatic, intelligent and digital twinization, gives typical scenarios, and puts forward suggestions on the evolution path of telco’s cloud-network convergence orchestrator and scheduler system.