Introduction

Telecom network services are the backbone of the modern digital economy, enabling voice, data, and video communication across the globe. However, the traditional telecom network infrastructure is facing several challenges, such as equipment obsolescence, high capital and operational costs, lack of scalability and flexibility, and increasing demand for bandwidth and low latency. To overcome these challenges, telecom network services are undergoing a transformation, moving from hardware-based to software-based solutions, leveraging cloud computing and artificial intelligence (AI) technologies. In this article, we will explore the history, trends, challenges, benefits, and use cases of telecom network services powered by cloud and AI / Gen AI based solutions, and how service providers like Eviden can offer innovative and value-added offerings in this space.

Brief history of Telecom network services and their migration to cloud based setup

Telecommunication services have evolved from the analog telephony and radio systems of the 19th and 20th centuries to the digital fixed, cellular and broadband networks of the 21st century, from circuit-switched to packet-switched networks, from 2G to 5G, and from centralized to distributed architectures. However, the traditional telecom network services are still based on proprietary hardware and software, which limit their ability to adapt to changing customer needs and market dynamics.

Evolution of Telecom Industry Technologies

The migration of telecom services to the cloud-based setup occurred in several phases, starting with the introduction of stored program control exchanges in the 1950s which introduced software to control telephone switching. Subsequently, the introduction of voice over IP began in the 1990s after the introduction of the public internet. Soon after telephone switching was implemented on commercial off the shelf servers. Then cloud based telecom services became possible with the emergence of cloud computing and software-defined networking (SDN) technologies, which enabled the virtualization and orchestration of network functions and resources. The cloud-based setup also enables the telecom operators to offer more flexible and customized services to their customers, as well as to leverage the cloud computing capabilities of data analytics, AI, and machine learning (ML).

Since the introduction of VoIP several over-the-top providers have flourished, such as Twillio, Ring Central, Vonage, and 8×8. Over-the-top means they run independently over broadband connections, in parallel to the Public Switched Telephone Network (PSTN). However, several countries have announced the “Sunset of the PSTN” including the US Federal Communications Commission mandate for 2025, and the UK Openreach plan also for 2025. From that point, the fixed line providers will offer VoIP as the only telephony option.

Telecom network services are migrating to cloud based setup, which offers several advantages, such as:

• Eliminate obsolete equipment from the network and replace it with the latest cloud based equipment.
• Reduced capital and operational expenses, as network functions are virtualized and run on commodity servers in the cloud.
• Increased scalability and flexibility, as network resources can be dynamically allocated and optimized according to the traffic and service requirements.
• Enhanced innovation and agility, as network services can be deployed and updated faster and easier, using cloud-native development and DevOps practices.
• Improved customer experience and satisfaction, as network services can offer higher quality, reliability, security, and personalization, using cloud-based analytics and AI

Telecom industry standards that support fixed networks and mobile 5G and 6G networks

The new telecom standards specify the development and deployment of fixed and mobile networks (5G and evolution to 6G), which require new architectures, protocols, and technologies to support the diverse and demanding requirements of these services. Telecom standards are defined by:

• The 3rd Generation Partnership Project (3GPP) is the leading standardization body for mobile communications. In 1999, 3GPP introduced standards for 3G which included both circuit switched mobile (Voice over LTE), and the IMS (IP Multimedia Subsystem) which was the first introduction of VoIP in mobile networks. IMS is based on the SIP Protocol defined by the IETF. 3GPP 5G standards have defined that 5G voice/video call services must use IMS and SIP. In addition, 3GPP has released specifications for the 5G New Radio (NR), which defines the air interface and radio access network (RAN) for 5G; the 5G Core Network (5GC), which defines the core network functions and services for 5G; and the 5G System Architecture (5GS), which defines the overall system architecture and interfaces for 5G.
• The Next Generation Mobile Networks Alliance (NGMN www.ngmn.org) is a sub-group of 3GPP which has published specifications for 5G and 6G which define the migration of the mobile network core to the cloud, and the use of AI/ML in the Radio Access Network to switch off cell radios during periods of low traffic to reduce energy consumption, and as a result reduce carbon, making the network greener.
• The Internet Engineering Task Force (IETF), which is the leading standardization body for internet protocols, has developed several standards that are leveraged by telecom networks, such as Session Initiation Protocol (SIP), which is the de-facto standard for VoIP signaling; Real Time Protocol (RTP) which defines how to transport voice over IP; SCTP and QUIC which are new transport protocols that provide faster and more reliable connections for 5G;
• The European Telecommunications Standards Institute (ETSI) and the American National Standards Institute (ANSI) are the primary standardization bodies for telecommunications. They released standards that formed the foundation of telecom networks such as SS7 (Signaling System No7) and ISDN (Integration Services Digital Network, and have established several standards for the 5G networks, such as the Multi-access Edge Computing (MEC), which is a framework that enables the deployment of computing and storage resources at the edge of the network, closer to the end-users and devices, for 5G; the Network Functions Virtualization (NFV), which is a framework that enables the virtualization and orchestration of network functions, such as firewalls, routers, or gateways, for 5G; and the Zero-touch network and Service Management (ZSM), which is a framework that enables the automation and optimization of network and service management for 5G.

The adoption of cloud-based setup and the new telecom industry standards will enable the deployment of 5G and 6G networks, which will offer unprecedented levels of speed, capacity, low latency, and connectivity, and enable a wide range of new applications and services, such as IoT (Internet of Things), AI/ML (Artificial Intelligence and Machine Learning), Cloud Gaming, Cloud Robotics and Cloud Virtual Reality and Artificial Reality.

Satellite Services

In a recent announcement Starlink, a subsidiary of Elon Musk’s SpaceX, plans to offer satellite-based mobile text and voice almost anywhere in the world using its low-earth orbit satellites that are already deployed. Starlink started out offering high-speed internet over satellite, which is especially valuable for low-population areas where broadband and mobile are not offered.

The latest announcement for Startlink is that it will offer text messaging to mobile phones in 2024, and voice in 2025. This will be offered to un-modified 4G LTE (Long Term Evolution) cellphones and later 5G/6G. It effectively makes the satellite network look like any other cell site. Starlink will partner with T-Mobile for mobile network services, including handing off voice calls for completion in the existing network. However, there is no reason why Starlink could not implement its own voice calling using VoIP, should it decide to do so in the future.

Adoption of AI, ML and Gen AI in Telecom industry

• AI, ML and Gen AI are the key technologies that will enable the transformation of telecom network services powered by cloud and AI / Gen AI based solutions. AI, ML and Gen AI can be applied to various aspects of telecom network services, such as:
• Optimal Radio Access Network operation, by turning off cell radios during periods of low traffic to reduce energy consumption, and thus reduce carbon emissions.
• Network planning and design, which involves optimizing the network topology, configuration, and parameters, to meet the service level agreements and quality of service requirements, and to minimize the network costs and risks.
• Network operation and maintenance, which involves monitoring, diagnosing, and troubleshooting the network performance, faults, and anomalies, and performing preventive and corrective actions, to ensure the network availability and reliability.
• Network security and privacy, which involves detecting, preventing, and mitigating the network attacks, threats, and vulnerabilities, and protecting the network data and assets, to ensure the network integrity and confidentiality.
• Network slicing and orchestration, which involves creating and managing multiple virtual and isolated network segments, each with distinctive characteristics and capabilities, and allocating and optimizing the network resources and functions, to meet the diverse and dynamic needs of different customers and applications.
• Network service and application provisioning, which involves delivering and managing the network services and applications, such as voice, data, video, IoT, AI, gaming, VR/AR, and robotics, and ensuring their quality, performance, and user satisfaction.
• AI, which can enable the optimization and coordination of satellite and terrestrial networks, and provide intelligent and adaptive network management and operation, to improve the network efficiency and performance.
• Gen AI, which can enable the creation and evolution of new network architectures and protocols, and provide innovative and customized network solutions, to meet the specific and changing needs of different customers and applications.

Telecom Industry Cloud

“Industry Cloud” is a trend discussed by analysts, such as Gartner. It states that the current general purpose hyperscaler clouds will evolve into several industry specific cloud variants, which address the specific needs of the target industry.  Industry Clouds will require close cooperation and three-way partnerships between the hyperscalers, global system integrators, and the relevant industry bodies.

A perfect example of the need for Industry Cloud is for the Telecoms industry. For example, the NGMN specifications of 5G and 6G specify that the mobile core network equipment will migrate to the central clouds of the hyperscalers. In addition, the mobile cell sites will evolve to use commercial off the shelf servers, under the control of the central cloud, in a typical edge computing architecture, using Kubernetes and GCP Anthos/Distributed Cloud Edge, AWS Outposts or Azure Arc/Azure Stack Edge.

NGMN also specifies the use of AI/ML running at the edge to turn off the wireless radios during periods of low traffic, this reducing energy usage, and making the solution more sustainable.

Industry Cloud should address the challenges of migrating telecom equipment to the cloud. The main challenge is to ensure that the new cloud deployments provide “Carrier Grade” reliability and availability. To be deployed in carrier networks, all equipment needs to support 99.999% availability (also referred to as five nines). This means that the equipment may be unavailable due to outages or maintenance for no more than 5.26 minutes per year. In addition, it is standard that all telecom equipment preserves stable calls (calls that are established) during switchover to standby. The three hyperscalers GCP, AWS and Azure typically advertise 99.95% (Three and a half nines) availability for their data centers and regions. This means that to achieve 99.999% availability it is necessary to use more than one region.

Geographic Redundancy is another requirement of availability that grew out of the migration of telecom to commercial servers and subsequently the cloud. Geo-redundancy means that service is provided from multiple locations (two or more) that are separated geographically. This is to ensure that service continues even if one of the locations becomes unavailable due to natural disasters, such as floods or hurricanes, or simple power outages.

The requirement to implement telecom workloads in multiple regions also presents some issues. The cloud is optimized for web-based protocols, HTTP and HTTPS (ports 80 and 443). Global IP load balancing can route traffic to multiple regions based on availability, but this only works for HTTP and HTTPS. Since telecom protocols use different ports (eg port 5060 for SIP) this traffic cannot be globally load balanced, and other techniques have to be used such as DNS to provide multiple IP addresses.

Another issue of multi-region deployments is data replication. Telecom Data includes subscriber profiles, and call state which is necessary for the saving of stable calls. Synchronous data replication is subject to the limitation of less than 5 ms round trip time, which limits the physical distance between nodes to approximately 300 km. This 5 ms limitation applies to both Synchronous SAN replication, and SQL database replication. As a result, the hyperscalers support Cloud SQL replication between zones of the same region, but usually not between regions. There are some exceptions; for example GCP Cloud Spanner databases can be replicated between two regions which are “close” together, but not between distant regions. Similarly, in memory databases such as Redis might be better suited to call state, but these only support replication within the same region.

An alternative is to use asynchronous replication between regions, such as is supported by AWS RDS, or to use NoSQL databases which use asynchronous replication. However, for call states these might not be sufficient, since consistency is critical.

Summary

Telecom networks both fixed and mobile are migrating to cloud based architectures, due to multiple benefits from reduced costs, improved efficiency and flexibility, faster innovation, and the benefits of cloud technologies such as AI and ML.

Eviden provides expertise in all three hyperscalers, with 19,500 certified cloud practitioners. In addition, Eviden has many experts in Telecom technologies. As such, Eviden is well positioned to provide advisory and implementation services as part of a Telecom Industry Cloud offering. Eviden will work with the hyperscalers and Telecom industry bodies to recommend best practices and products for migration of telecom equipment to the cloud, taking into account the specific requirements of Telecom. Once the optimum cloud configurations are defined, these can be implemented by Infrastructure as Code modules which speed the deployment in the cloud.