Service-partner-oriented Network Slicing

1. Introduction

In the fifth-generation mobile communications (5G) era, it is expected that various new services will be created by taking advantage of the features of 5G such as high-capacity broadband, massive session connectivity, and ultralow-latency and high-quality communications. To realize these new services, a variety of networks will be required to meet various service requirements. The challenge is to provide networks quickly and flexibly in response to such demands.

2. Network slicing technology

Network slicing technology enables the simultaneous construction and operation of multiple virtual networks called slices having different requirements on a common physical infrastructure. Physical facilities are managed as resources that can be virtually partitioned, and these resources can be freely combined to create the virtual network needed. Conventional communications networks use expensive specialized equipment that takes time to build. In contrast, with slices, various services can be promptly provided by switching settings using general purpose equipment that is relatively inexpensive. This technology is expected to support networks of the 5G era. At NTT, our aim is to use this technology to quickly provide our service partners with the networks they need.

3. Slice gateway technology

The provision of networks for service partners with diverse requirements requires end-to-end slices, each of which can maintain a fixed level of communications quality on an end-to-end basis. One problem here is the construction and operation of end-to-end slices that straddle multiple provider networks managed under various rules. To solve this problem, NTT is proposing an architecture that deploys slice gateways (SLGs) at connection points between provider networks (Fig. 1). An SLG provides functions needed on a slice data plane such as protocol conversion, traffic allocation, and inter-slice isolation. The SLG deployed at the connection point between the networks can operate the slice across multiple networks with different specifications by performing appropriate conversion according to the specifications of each network.

Fig. 1. Overview of proposed architecture.

We are currently installing SLGs using open source software (OSS) on the NetroSpherePIT [1] test platform and conducting trials. A slice can be constructed using any protocol compatible with the specifications of the infrastructure network. We are investigating the construction of slices using the new technology SRv6 (Segment Routing Internet Protocol version 6). We are also constructing and evaluating end-to-end slices with latency guarantee [2] as an added function with the aim of providing an even better user experience. Finally, we have also developed a prototype GUI (graphical user interface) for constructing and switching slices and displaying slice operation status through telemetry technology. The research and development (R&D) results described above were presented at NTT R&D Forum 2018 Autumn held in November 2018. At this exhibit, service partners were able to experience slice control for themselves.

4. Future outlook

NTT is moving forward on refining network slicing and SLG technologies through test bed trials with an eye to commercialization. We are currently promoting global standardization with the aim of including these technologies in commercial products and de facto OSS.

References

[1]	T. Okutani, A. Kawabata, T. Kotani, T. Yamada, and M. Maruyama, “NetroSpherePIT: Demonstrations to Accelerate the Adoption of NetroSphere,” NTT Technical Review, Vol. 14, No. 10, 2016. https://www.ntt-review.jp/archive/ntttechnical.php?contents=ntr201610fa2.html
[2]	T. Fukui, Y. Sakaue, and K. Minami, “Guaranteed Transmission within Maximum Allowable Network Latency for Enhanced User Experience in Two-way Communication Applications,” NTT Technical Review, Vol. 17, No. 6, pp. 20–22, 2019. https://www.ntt-review.jp/archive/ntttechnical.php?contents=ntr201906fa6.html