Joint Placement and Allocation of Virtual Network Functions with Budget and Capacity Constraints

Abstract

With the advent of Network Function Virtualization (NFV), network services that traditionally run on proprietary dedicated hardware can now be realized using Virtual Network Functions (VNFs) that are hosted on general-purpose commodity hardware. This new network paradigm offers a great flexibility to Internet service providers (ISPs) for efficiently operating their networks (collecting network statistics, enforcing management policies, etc.). However, introducing NFV requires an investment to deploy VNFs at certain network nodes (called VNF-nodes), which has to account for practical constraints such as the deployment budget and the VNF-node capacity. To that end, it is important to design a joint VNF-nodes placement and capacity allocation algorithm that can maximize the total amount of network flows that are fully processed by the VNF-nodes while respecting such practical constraints. In contrast to most prior work that often neglects either the budget constraint or the capacity constraint, we explicitly consider both of them. We prove that accounting for these constraints introduces several new challenges. Specifically, we prove that the studied problem is not only NP-hard but also non-submodular. To address these challenges, we introduce a novel relaxation method such that the objective function of the relaxed placement subproblem becomes submodular. Leveraging this useful submodular property, we propose two algorithms that achieve an approximation ratio of 1 2 (1 − 1/e) and 1 3 (1 − 1/e) for the original non-relaxed problem, respectively. Finally, we corroborate the effectiveness of the proposed algorithms through extensive evaluations using both trace-driven simulations and simulations based on synthesized network settings.

Online Joint Placement and Allocation of Virtual Network Functions With Heterogeneous Servers

“…However, a hidden assumption in the above work is that any virtual server in the given network has unlimited potential to activate VNF instances, while it is not always the truth. In the latter work of Sallam and Ji [16], they remove this assumption and make a new one that once a virtual server is activated to be a VNF node, it has a fixed limited computing resources to serve flow demand. Each virtual server i in the network has its specific amount c i of potential capacity and specific cost b i to be activated.…”

Placement and Allocation of Virtual Network Functions: Multi-dimensional Case

“…The placement problem has been considered in different domains such as NFV (e.g., [5]), SDN (e.g., [3]), and edge cloud computing (e.g., [12]). In NFV, several studies (e.g., [4], [13], [14]) consider the placement of a minimum number of VNF instances to cover all flows.…”

“…Again, this work does not consider multiple network functions or budget constraint. The closest work to ours is [5], where budget, resource, and fully processed flow constraints are considered, but it only considers one type of network functions and only a single type of resource.…”

“…Most of the previous work either does not have a limited budget (e.g., [4]) or relaxes the resources constraint (e.g., [3]). In [5], both the budget and resources constraints are considered, along with fully processed flows requirement. However, they consider a special case of the VPRA problem with the following characteristics: a) there is only one type of resource; b) all flows require the same network function; and c) the network function requires one unit of resource to process each unit of flows (we refer to this setting as basic-VPRA).…”

“…Even under such a simplified setting, the basic-VPRA is already quite challenging. It is shown in [5] that the problem is not only NP-hard but also non-submodular (a property that generally leads to efficient solutions for similar problems (e.g., [3])). In this work, we take one step further and extend the basic-VPRA problem to consider the setting with multiple network functions, multiple resources, and heterogeneous resource requirements.…”