Telecom Infra Project—Its Structure and Activities

1. Background

In the technological fields involving datacenters and datacenter interconnections, de facto standards are widely used. They come from products and specifications developed by organizations such as the Open Networking Foundation as well as open source projects. What is characteristic is that these standards are developed through making the designs and interfaces open. In many cases, the resulting products are also open source. The trend of openness is now seen in not only the development of software but also of hardware. For example, the Open Compute Project (OCP) pushed through openness in the specifications of servers and racks used in datacenters, leading to a reduction in the purchasing cost as well as in the power consumption. In this approach, functions and components were unbundled. It is this disaggregation that made it possible for users to choose only the parts necessary to them, eliminating others.

The trend of openness and disaggregation is now beginning to expand to telecom networks. Telecom Infra Project (TIP) was established with the leadership of Facebook, one of the leading companies of OCP, and with the idea that the same approach as OCP’s could be applied in telecom networks.

2. Overview of TIP

Since TIP started its activities in February 2016, the number of member companies has increased from several dozen to over 500. NTT joined TIP in 2017. The Board of Directors includes people from telecom operators such as Deutsche Telekom (DT) and SK Telecom, which have held the position since the foundation of the organization, and from Telefónica, Vodafone, and British Telecom (BT). This demonstrates that European operators are actively involved in the activities. Many operators from other regions such as India, South Africa, and Brazil are also participating. For these operators whose service areas include places where communications infrastructure has not been well established, the open-sourced products and the know-how arising from TIP would serve as a solution to address the issue adequately and cost-effectively.

Manufacturers of white box equipment, which use merchant chips and non-proprietary operating systems, and chip vendors are among those who have expanded their presence in concert with the markets of these products. There are also other kinds of participants such as Internet service providers and system integrators, including startup companies. There is a great deal of interest in the latest technical trends and expectations for their implementation among the wide variety of participants.

The organizational structure of TIP is shown in Fig. 1. Project Groups (PGs), responsible for creating specifications or developing products, are established after their charters are accepted by the Board of Directors and the Technical Committee. Although some PGs have been dissolved, the number of PGs has increased since TIP was founded, showing flexibility similar to that of other organizations in the field of datacenter technologies.

Fig. 1. Organizational structure of TIP.

PGs are classified by their scopes into three technical fields: Access Projects, Backhaul Projects, and Core and Management Projects. The scopes of individual PGs are written in their charters along with goals and deliverables. Products developed by a PG are usually tested in a Community Lab. There are currently six Community Labs, two in the USA and one each in Germany, South Korea, Brazil, and India. Each Community Lab is hosted and operated by a specific company with a policy of allowing the participation of member companies. As is evident in many successful open source projects, creating open communities is recognized as a key to advancing development.

TIP Ecosystem Acceleration Centers (TEACs) provide startup companies with an environment for product development and tests. Currently, BT, DT, Orange, and SK Telecom each host a TEAC.

3. Access Projects

The activities of the PGs whose scope is related to the Access field are listed in Table 1 [1]. These PGs aim to make Internet access easier by innovating access network technologies. Reducing infrastructure costs and creating new technologies that are applicable where existing ones cannot be easily used are among the examples of the PGs’ goals.

Table 1. Access Projects.

The OpenCellular PG is working to produce base stations of Long-Term Evolution (LTE) cellular networks. While open source software is employed for the management of the base stations, chips and other hardware are developed by the participant companies, and their design has been made public as open source hardware. This PG is focusing on the existing mobile technology, rather than the emerging fifth-generation (5G) mobile technology, and it has brought openness into its products, representing the TIP’s concept. The products have been used in field trials in India, Pakistan, and other African and South American countries.

The activities of many of the other Access PGs are also related to mobile communications infrastructure. The vRAN Fronthaul PG is tackling issues on the mobile fronthaul, which is responsible for the connection between antennas and baseband signal processing units of base stations. In developing countries, optical fibers cannot always be used as transmission media due to costs or environmental problems. The PG is addressing this issue by studying how to implement fronthaul transmission using twisted pair cables, coaxial cables, or other non-ideal media.

A power supply can also be an impediment in deploying infrastructure that costs or environmental problems impose. The Power and Connectivity PG aims to establish technologies and business models to deploy power supply infrastructure in rural areas with lower costs, inviting power companies and battery makers as members.

The OpenRAN PG’s scope is similar to that of the OpenCellular PG’s in that both are aimed at developing LTE base stations. Their focus, however, is on software implementation and its programmability. Because the functionalities are programmed to work with general purpose processors, flexible and timely functional updates are possible.

4. Backhaul Projects

A rapid increase in the amount of communications traffic is being seen in many countries and is a common issue for telecom operators worldwide. Against this background, PGs related to the Backhaul field are trying to innovate technologies in trunk and backbone networks. The activities of the PGs focusing on this field are summarized in Table 2 [1].

Table 2. Backhaul Projects.

The Open Optical Packet Transport PG is targeting innovative optical fiber transmission techniques. They have released two equipment prototypes named Voyager and Cassini. As well as being optical transponders with integrated packet switching functions, they are characterized by the use of nonproprietary components. In Cassini, for example, a merchant digital signal processor is employed to implement coherent optical transmission, which NTT’s laboratory has developed with its partners and is now sold by NTT Electronics. The operating system employed in it is also a commercial, nonproprietary one.

This type of equipment, referred to as white box, enables multi-vendor implementation in such a way as to enable specific parts to be replaced with those of another vendor. Although the usage of Voyager or Cassini seems to be limited to a point-to-point interconnection between datacenters, the trend of disaggregation that they embody is likely to penetrate into more complicated optical transmission equipment.

In the field of wireless transmission, the Millimeter Wave Network PG is working to establish an infrastructure technology using 60-GHz radio waves. The frequency falls in an unlicensed band in many countries, including Japan and other Asian countries, Europe, and North and South American countries, reducing hurdles for its use. The wireless infrastructure can also be installed more easily than optical fiber. With these advantages, the PG is planning a field trial to obtain practical expertise to apply the technology in city areas where capacity demands tend to increase in a short period.

5. Core and Management Projects

The goal of PGs focusing on the Core and Management field is to innovate network operations. Their activities are summarized in Table 3 [1].

Table 3. Core and Management Projects.

The End-to-End Network Slicing PG deals with virtualized networks, or network slices, that are built on physical infrastructure. Their goal is to achieve operation that involves configuration and management of network slices spanning multiple operators’ domains. To that end, they study interface specifications that enable cooperation of multiple orchestrators, each of which manages network slices and their constituent physical resources in a specific domain. Their scope is aligned with emerging 5G mobile technologies.

The Artificial Intelligence and Applied Machine Learning PG is working to introduce artificial intelligence (AI) in network operations. They study how to streamline operation processes or how to optimize networks by utilizing AI where human judgment currently intervenes. Although their work is in an early phase, it could attract attention as a new approach for introducing innovation.

6. Outlook

The advancement of processor chips’ capabilities and software engineering techniques and the trend of openness have brought various innovations, disaggregation being an example. Disaggregation, or unbundling functions or parts that used to be integrated, sometimes as a black box, was first introduced in datacenters.

Because telecom networks contain a wide variety of equipment in terms of both functionality and size, it is not likely that the same techniques will be directly introduced into telecom networks immediately. However, it is almost certain that the approach will have a significant influence on the research and development of telecom networks.

NTT has been participating in several PGs of TIP not just to observe their activities but also to make contributions. For example, in the Open Optical Packet Transport PG, we have proposed a packet switching and forwarding architecture called Multi-Service Fabric (MSF) [2, 3] that enables highly reliable, multi-vendor implementations. Through the activities at TIP, we hope to keep engaging ourselves in the newest technical trends and use them in our research and development.

References

[1]	Excerpts from TIP website, https://telecominfraproject.com/
[2]	K. Takahashi, H. Yoshioka, K. Ono, and T. Iwai, “Promoting the MSF Architecture for Flexible Networks,” NTT Technical Review, Vol. 14, No. 10, 2016. https://www.ntt-review.jp/archive/ntttechnical.php?contents=ntr201610fa6.html
[3]	Multi-Service Fabric (MSF), https://github.com/multi-service-fabric/msf

Trademark notes

All brand names, product names, and company names that appear in this article are trademarks or registered trademarks of their respective owners.