Closed ashwinprasadme closed 5 years ago
• This paper addresses the information exchange management requirements of NFV environments by introducing virtual infrastructure information service (VIS) as a service facility for a MANO framework.
• This helps us in addressing how to perform state management
• This paper is about the placement of number of NFVOs and VNFMs in MANO while addressing it’s scalability.
• This addresses the number of NFVOs, VNFMs needed for a distributed NFVI-PoPs, which helps our research to find the optimal number of MANO required in a system.
This paper investigates the hierarchical approach for service placement, helping our research in instantiating MANO at different levels.
This paper helps us to know the service management for NFV services as a microservice in a multiple MANO environment.
Section V of this paper gives an overview of the monitoring mechanisms of an NFV system.
@mehraghdam We have completed the initial reading and we would like to discuss with you before taking it forward.
Sure.
A survey: Control plane scalability issues and approaches in Software-Defined Networking (SDN)
Reading this paper to get a general idea of how the scalability problem is approached.
Scalability in general can mean many things according to context.
Scalability involves trade-offs, not an independant problem.
Authors identify primary concern as "SDN controller" performance. What will be our primary concern?
Prominent metrics considered in the paper is control plane throughput and latency. We need to identify this in our case.
Logically centralized architecture, prone to scalability issues.
Authors identify 3 contributors to SDN scalability issues. like separation of control and data planes, quantity of requests and communication delay. We need to establish the major contributors in our case.
Authors review the scalability issues by classifying them into categories like Topology and Mechanisms related approaches.
Topology based study here is interesting in our case, we can learn more from Section 5 here.
Mechanism approaches are about using optimizations, like multi-thread to achive better results.
Important thing is the distrubution of state and consistency between them. This is also something relavent to us.
Management and Orchestration Challenges in Network Functions Virtualization-
Reading this to understand the scalability challenges of MANO
Section "Challenges and Research Opportunities" lists some challenges which are of interest,
Section "Distributed Management" mentions the scalability challenge which arraises as a result of centralization
Optimising Microservice-based Reliable NFV Management & Orchestration Architectures
Interesting paper that investigate similar research problems as ours
Describes mechanisms that provide MANO with high availability and fault recovery.
Section 2 defends the choice of microservices in MANO frameworks by describing the workflow.
Authors discuss a solution to make MANO frameworks more reliable by maintaining multiple instances, distrubuting load and having state sharing between them. This option can be explored by us as well.
T-NOVA: An Open-Source MANO Stack for NFV Infrastructures
Introduction section gives a an overview of various operations expected from a MANO and it's challanges.
Describes why scalability is important in the monitoring framework.