A review on P4-Programmable data planes: Architecture, research efforts, and future directions

被引:35
作者
Kaur, Sukhveer [1 ]
Kumar, Krishan [1 ]
Aggarwal, Naveen [1 ]
机构
[1] Panjab Univ, Univ Inst Engn & Technol UIET, Chandigarh, India
关键词
P4; Programmable data plane; Software defined networking; OpenFlow; Control plane; Data plane; DDoS; Monitoring; Load balancing; SOFTWARE-DEFINED NETWORKS; SDN; SECURITY; PROGRAMMABILITY; SCALABILITY; TAXONOMY;
D O I
10.1016/j.comcom.2021.01.027
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Software Defined Networking (SDN) is a promising technology that provides flexibility, programmability, and network automation by shifting the network intelligence to the logically centralized controller. In SDN architecture, the controller maintains the global view of network topology; therefore the network management is efficient as compared to the traditional networks. Moreover, the cost of SDN devices is less due to the use of open source network operating system instead of proprietary and vendor-specific software. In spite of the flexibility offered by the SDN, OpenFlow enabled switches have a fixed behavior as specified in the datasheet of switch ASIC. They recognize fixed set of header fields according to the support of OpenFlow version. In addition, SDN is suffering from scalability and performance issues because SDN switches heavily dependent on the control plane to forward the packets which increases the data-control communication overhead. To resolve these issues, P4-Programmable data plane switches can be used. The analysis of review articles about SDN suggests insufficient focus on the data plane programmability. Therefore, this paper provides the comprehensive overview of domain-specific language (P4) for the programmability of the data plane. We have discussed the problems in the traditional SDN architecture and then, how these problems can be managed by the data plane programmability. Further, this study critically analyze the research articles based on the P4 programming language for network traffic monitoring, DDoS attack detection, load balancing, and packet aggregation and disaggregation. Finally, we have identified the research gaps and highlighted the open research challenges in the field of data plane programmability for the future directions.
引用
收藏
页码:109 / 129
页数:21
相关论文
共 104 条
  • [1] Detecting heavy flows in the SDN match and action model
    Afek, Yehuda
    Bremler-Barr, Anat
    Feibish, Shir Landau
    Schiff, Liron
    [J]. COMPUTER NETWORKS, 2018, 136 : 1 - 12
  • [2] Afek Y, 2017, IEEE INFOCOM SER
  • [3] Security in Software Defined Networks: A Survey
    Ahmad, Ijaz
    Namal, Suneth
    Ylianttila, Mika
    Gurtov, Andrei
    [J]. IEEE COMMUNICATIONS SURVEYS AND TUTORIALS, 2015, 17 (04): : 2317 - 2346
  • [4] A roadmap for traffic engineering in SDN-OpenFlow networks
    Akyildiz, Ian F.
    Lee, Ahyoung
    Wang, Pu
    Luo, Min
    Chou, Wu
    [J]. COMPUTER NETWORKS, 2014, 71 : 1 - 30
  • [5] CONGA: Distributed Congestion-Aware Load Balancing for Datacenters
    Alizadeh, Mohammad
    Edsall, Tom
    Dharmapurikar, Sarang
    Vaidyanathan, Ramanan
    Chu, Kevin
    Fingerhut, Andy
    Vinh The Lam
    Matus, Francis
    Pan, Rong
    Yadav, Navindra
    Varghese, George
    [J]. ACM SIGCOMM COMPUTER COMMUNICATION REVIEW, 2014, 44 (04) : 503 - 514
  • [6] Security of Software Defined Networks: A survey
    Alsmadr, Izzat
    Xu, Dianxiang
    [J]. COMPUTERS & SECURITY, 2015, 53 : 79 - 108
  • [7] Hybrid SDN Networks: A Survey of Existing Approaches
    Amin, Rashid
    Reisslein, Martin
    Shah, Nadir
    [J]. IEEE COMMUNICATIONS SURVEYS AND TUTORIALS, 2018, 20 (04): : 3259 - 3306
  • [8] [Anonymous], 2017, Complete Visual Networking Index, V12, P749
  • [9] [Anonymous], 2015, BEHAV MODEL BMV2
  • [10] [Anonymous], The World's Fastest Most Programmable Networks