Demystifying the Performance Interference of Co-Located Virtual Network Functions

Zeng, Chaobing; Liu, Fangming; Chen, Shutong; Weixiang, Jiang; Li, Miao

doi:10.1109/infocom.2018.8486246

Cited by 48 publications

(12 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…can be provided by one or more service function chains (SFCs), i.e., a concatenation of appropriate VNFs that must be crossed by the traffic associated to that specific service. The main weakness of NFV is the hard-to-predict performance due to resource sharing of hardware among different functions, especially concerning the processing [3].…”

Section: Introductionmentioning

confidence: 99%

Impact of Processing-Resource Sharing on the Placement of Chained Virtual Network Functions

Savi

Tornatore

Verticale

2021

IEEE Trans. Cloud Comput.

View full text Add to dashboard Cite

Network Function Virtualization (NFV) provides higher flexibility for network operators and reduces the complexity in network service deployment. Using NFV, Virtual Network Functions (VNF) can be located in various network nodes and chained together in a Service Function Chain (SFC) to provide a specific service. Consolidating multiple VNFs in a smaller number of locations would allow decreasing capital expenditures. However, excessive consolidation of VNFs might cause additional latency penalties due to processing-resource sharing, and this is undesirable, as SFCs are bounded by service-specific latency requirements. In this paper, we identify two different types of penalties (referred as "costs") related to the processingresource sharing among multiple VNFs: the context switching costs and the upscaling costs. Context switching costs arise when multiple CPU processes (e.g., supporting different VNFs) share the same CPU and thus repeated loading/saving of their context is required. Upscaling costs are incurred by VNFs requiring multi-core implementations, since they suffer a penalty due to the load-balancing needs among CPU cores. These costs affect how the chained VNFs are placed in the network to meet the performance requirement of the SFCs. We evaluate their impact while considering SFCs with different bandwidth and latency requirements in a scenario of VNF consolidation.

show abstract

Section: Introductionmentioning

confidence: 99%

Impact of Processing-Resource Sharing on the Placement of Chained Virtual Network Functions

Savi

Tornatore

Verticale

2021

IEEE Trans. Cloud Comput.

View full text Add to dashboard Cite

show abstract

“…LLC interference in NFV environments was not widely studied yet, nevertheless there are some works which study LLC interference and aim to reduce LLC interference [1], [5], [6], [7]. NFV also emerged from cloud compute concepts, but LLC interference scheduling is studied more in depth with respect to compute cloud environments.…”

Section: Related Workmentioning

confidence: 99%

“…Recently authors of [1] studied the interference effects of co-located VNFs in depth. They consider contention of network I/O bandwidth, CPU, memory and cache.…”

Section: B Llc Interference In Nfv Environmentsmentioning

confidence: 99%

“…if some data is only used for the initialization of the VNF. Further, the access pattern depends also on the network condition, e.g., an IDS serving highly diverse traffic has a different access pattern than an IDS that only filters a single connection [1].…”

Section: Memory Access Modelmentioning

confidence: 99%

“…On the other hand, this pattern causes interference between the consolidated VNFs at different places in the shared system. [1] shows that VNFs are competing for network IO bandwidth, CPU, cache and memory. As a result, co-location can half the throughput of one VNF compared to the case the when VNF is running alone due to interference.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Towards Reducing Last-Level-Cache Interference of Co-Located Virtual Network Functions

Durner

Sieber

Kellerer

2019

2019 28th International Conference on Computer Communication and Networks (ICCCN)

View full text Add to dashboard Cite

Network Function Virtualization (NFV) aims to virtualize compute resources for packet processing in order to gain flexibility and reduce costs. In order to increase the resource utilization, multiple VNFs are co-located on one single server. Current virtualization techniques do not fully isolate all resources, thus co-location of VNFs causes interference effects. It has been shown that these interference effects can degrade the performance of Virtualized Network Functions (VNFs) in terms of throughput and delay severely. In this work we aim to gather the potential that lies in reduction of the interference due to the shared Last Level Cache (LLC). CPU caches are used to improve the access times to memory that is needed regularly for the execution of a program. Intel Cache Allocation Technology (CAT) provides the means to allocate the cache and isolate VNFs from each other. The results show that the scheduler can decrease the CPU utilization by up to 20%. We can show which factors influence the gain of LLC scheduling in NFV deployments. In order to show this we propose a scheduler which optimally allocates the LLC in order to reduce the maximum CPU utilization of all cores.

show abstract

Modeling on virtual network embedding using reinforcement learning

Wang

Zheng

et al. 2020

Concurrency and Computation

View full text Add to dashboard Cite

It is well known that virtual network (VN) embedding (VNE) aims to solve how to efficiently allocate physical resources to a VN. However, this issue has been proved to be an NP-hard problem. Besides, as most of the existing approaches are based on heuristic algorithms, which is easy to fall into local optimal. To address the challenge, we formalize the problem as a mixed integer programming problem and propose a novel VNE method based on reinforcement learning in this article. And to solve the problem, we introduce a pointer network to generate virtual node mapping strategies through an attention mechanism, and design a reward function related to link resource consumption to build the connection between node mapping and link mapping stages of VNE. In addition, we present a policy gradient optimization mechanism to leverage the reward information obtained from the sampled solutions, and design an active search based process to automatically update the parameters of the neural network and to obtain near-optimal embedding solution. The experimental results show that the proposed method can improve the performance in average physical node utilization and long-term revenue to cost ratio comparing than that of the existing models. K E Y W O R D S attention, pointer network, policy gradient, reinforcement learning, virtual network embedding 1 INTRODUCTION Network virtualization technology can allow multiple heterogeneous network architectures to coexist on a shared physical substrate network (SN). 1 In network virtualization, the role of traditional Internet service providers (ISPs) is separated into two independent parts: infrastructure providers (InPs), who manage the physical infrastructure, and service providers (SPs), who provide services by renting virtual network (VN) using resources from InPs. 2 In such a way, enterprises are increasingly deploying services 3 to the Cloud with the development of network function virtualization (NFV) 4 and software defined networking (SDN). 5 The major challenges of network virtualization are how to effectively allocate SN resources to VNs as well as increase the utilization and revenue of InPs, which is known as the virtual network embedding (VNE) problem. 6 The VNE problem mainly includes two parts: node mapping stage and link mapping stage. It has been proven to be NP-hard. 7 Many exact and heuristic algorithms 8-13 have been proposed to solve the VNE problem in recent years. Although exact solutions can achieve high resource utilization rate of the SN, they require exponential time-consuming and therefore are not practical. In addition, most of the heuristic algorithms for VNE consider link mapping and node mapping separately. They did not consider the inner practical connection between the two stages and may result in suboptimal solutions. Furthermore, they are unable to leverage historical experience to optimize themselves automatically.

show abstract

Demystifying the Performance Interference of Co-Located Virtual Network Functions

Cited by 48 publications

References 21 publications

Impact of Processing-Resource Sharing on the Placement of Chained Virtual Network Functions

Impact of Processing-Resource Sharing on the Placement of Chained Virtual Network Functions

Towards Reducing Last-Level-Cache Interference of Co-Located Virtual Network Functions

Modeling on virtual network embedding using reinforcement learning

Contact Info

Product

Resources

About