دانلود رایگان ترجمه مقاله زنجیره سرویس پویا مبتنی بر سیاست در مجازی سازی عملکردهای شبکه (نشریه IEEE 2016)

زنجیره سرویس پویا مبتنی بر سیاست در مجازی سازی عملکردهای شبکه

Policy-Based Dynamic Service Chaining in Network Functions Virtualization

Abstract

Network Functions Virtualization (NFV) enables the rapid development, flexible management, and the dynamic placement of new, innovative Virtualized Network Functions (VNFs), such as load balancers, firewalls, and Intrusion Detection Systems (IDSes). Furthermore, NFV along with Software-Defined Networking (SDN) allows VNFs and physical middleboxes to be dynamically composed into service chaining graphs. Despite these benefits, service chaining graphs can be further improved through the use of techniques that have not been satisfactorily explored yet, such as Policy-Based Network Management (PBNM). In PBNM, policies can be written and triggered during runtime, thus supporting the dynamic (re)configuration of service graphs with minimal disruption. In this paper, we propose an approach to automatically design NFV service chaining graphs based on policies. These policies rule the forwarding of traffic and the construction of service chaining graphs. In our approach, service chaining graphs are enforced dynamically in the network during runtime. Finally, to assess its feasibility and generality, we create two different scenarios to demonstrate and discuss how our solution can be employed and its expected results.

1 Introduction

Network functions, such as load balancing, firewalls, and Intrusion Detection Systems (IDSes) are traditionally realized in physical devices often refereed to as middleboxes. Middleboxes tend to be proprietary and vendor-specific, and thus force network operators to learn about the peculiarities of middleboxes from different vendors, which is counterproductive. Also, physical middleboxes are not flexible enough to accommodate bursts of demand, which intrinsically hinders their scalability. Network Functions Virtualization (NFV) [1] is a novel technology that addresses the lack of flexibility of physical middleboxes. NFV proposes the use of Commercial Off-The-Shelf (COTS) hardware to host virtualized network services. With this approach, the capital expenditure (CAPEX) and operational expenditure (OPEX) can be significantly reduced. Also, with NFV, service provisioning can be easily scaled up and down according to network demands.

NFV allows the chaining of multiple Virtualized Network Functions (VNFs). Such VNF chaining enables network operators to dictate to which sequence of VNFs a packet should go through. The act of specifying the sequence of VNFs is called network service chaining [2]. Service chaining on current network infrastructures is statically defined and dependent on the network’s topology. This imposes a challenge to the operator when adding or removing services, considering that earlier technologies are difficult to redeploy [3]. With NFV and Software-Defined Networking (SDN) [4], this chaining can be performed dynamically. SDN decouples the control plane from the data plane, providing a global view of the network and a controller that performs traffic forwarding decisions [4]. With this separation, a controller can be implemented to steer the traffic dynamically during runtime. Therefore, service chaining can be easily adapted to the administrator’s need. This chaining is created from existent VNFs and middleboxes, thus using network resources efficiently [5].

Network operators have different needs according to the traffic of the networks that they manage. Also, network users do not necessarily need the same services (e.g., packets exchanged inside the enterprise’s network can pass through a simple firewall instead of a more sophisticated one). From these premises, a question emerges: how can the operator dynamically compose a set of VNFs to handle customized traffic flows? An approach to solving this problem is to use policies to govern the service chain of a flow. Policy-Based Network Management (PBNM) in computer networks is a concept widely applied and well-defined [6], but as long as the authors of this paper are aware of, its use in NFV service chaining has not been exploited.

In this paper, we present a PBNM solution to design and manage service chaining, where business-level operators can write Service Level Agreements (SLAs) to guide the building of service chaining graphs. We also introduce a Controlled Natural Language (CNL) to establish requirements and constraints for the writing of policies. Further, we discuss our solution’s feasibility and generality in two different scenarios. Our proposed solution can be employed in both homogeneous environments (VNFs only) and heterogeneous environments (composed of both VNFs and physical middleboxes).

This paper is structured as follows. In Section II, we review work related to this approach. Then, in Section III, the solution and associated architecture are described. In Section IV, two case studies are outlined and discussed. Finally, in Section V, we finish this paper with conclusions and future work.