How are network monitoring and optimization changing as 5G networks mature? RCR Wireless News reached out to Matt Beal, senior vice president of software development for Oracle Communications, for perspective on how it is helping its Communications Service Provider customers with the increasing complexity of 5G networks.
This Q&A was conducted via email and has been edited lightly for length.
How do you see network deployment, optimization and monitoring changing in 5G as opposed to 4G? What are some of the new moving parts that operators have to take into account with this technology shift, which includes cloud, edge and opening up the RAN as well?
Changes in network deployment, optimization and monitoring will be significant in 5G as opposed to 4G. This is being driven and enabled by the cloud native architecture of 5G. Cloud native architectures are operated at scale using more data- and insights-driven techniques fed by machine-learning algorithms. It also is enabling service providers to leverage open-source monitoring toolkits. This is not the case in most 4G networks which are, at best, rules-driven, leverage traditional low-volume monitoring tools, and rarely employ advanced machine learning.
This change is naturally taking place over time because carriers manage their networks as multigenerational networks. However, more advanced carriers will implement cloud native operating models and extend them to their 4G estates, especially where the 4G network embraces cloud native technologies both at the core and in the RAN. Great examples of operators leading the market in their approach to cloud native technology are Oracle customers Orange and DISH Wireless. To launch scalable, agile, and secure 5G core networks that can adapt and evolve to emerging enterprise and industry use cases, the operators are taking advantage of Oracle’s suite of cloud native network functions and 5G core technology.
With Open RAN coming into play as well as edge computing, operators are bringing on new vendors and ecosystem partners who come from cloud or IT as opposed to a telecom-specific background. What does the increased software-ization of network elements, and network elements and services operating on others’ cloud infrastructure, mean for network management, operations and assurance?
Software-ization and cloud-ification of network elements requires a massive shift in the mindset and capabilities of the operator. Disciplines like data management, infrastructure automation, and software lifecycle management must be adopted wholesale from cloud service providers. Carriers that seek to reinvent these disciplines or invent them from scratch can often ignore too many hard-won best practices, most of which have been codified in open-source tools and toolchains that companies like Oracle are now optimizing for carrier deployment.
The three most crucial changes that are emerging from the cloud-native transformation of the carrier network in 5G and OpenRAN/ORAN are:
- Automation: Velocity, scale, and quality will be dictated by the degree of automation achieved in the carrier network. Hyperscale cloud operators have achieved a very high degree of automation, one that drives from the service layer to the infrastructure layer, which is something carriers have never before delivered.
- Programmability: Network Exposure Function (NEF) and Network-as-a-Service strategies are emerging at a few carriers, which demonstrates there is finally a legitimate opportunity for carriers to control and manage their networks in software – and to extend control to their customers and the broader ecosystems they seek to create. Full automation is a mandatory pre-requisite to programmability, and arguably the single, largest reason many carriers have not yet achieved true programmability of the network estate. Limited programmability has existed previously, even in end-user services in the Class 5 switching domain, but this has never been broadly exposed or openly scaled.
- Visibility: The data complement of the NEF is the Network Data Analytics Function (NWDAF) which promises the first standardized function to grant access to the wealth of data that has been previously locked behind the implementations of 3GPP networks. With automation and programmability in hand, context and state can be understood from the insights generated by the NWDAF and other advanced analytics capabilities now being implemented in the 5G network. Fed into rich data stores and modern analytics models, carrier network, device, and customer data promise to generate insights, capabilities, and services never possible before.
These capabilities have been pioneered and proven by hyperscale cloud operators to create scale, performance, and services beyond what had been previously delivered in the IT or Telecommunications industries. Deploying these techniques in a similar way at the core of new carrier cloud networks will generate analogous benefit.
Who is integrating these new software-based networks? Do you see operators doing it themselves, turning to SIs or other partners? What are some of the new facets to integration in 5G networks? How does Oracle approach integration of its solutions into a multi-vendor 5G network environment?
Integration of the software-based network is being performed by a variety of players. In some cases, carriers are playing the role of prime integrator. In some they are tasking a system integrator and in others they are asking one of their network vendors to take on that responsibility. Practically every configuration can be found, depending on the current capabilities of the carrier, how they judge their ecosystem partners, and their assessment of their strategic imperative in undertaking the transformation. However, nearly every carrier understands this is as much a cultural transformation as a technical change. They must most formulate their vendor choices, their business case, and their integration strategy to drive their business and cultural change initiative.
In this environment, and as a best-of-breed provider of the most important network components, Oracle integrates flexibly into whatever integration strategy their carrier partner selects. In some cases, Oracle even extends its product capabilities to help facilitate the technical, process, or even cultural transformation as evidenced by its cloud native environment (CNE and vCNE), Automated Test Suite (ATS) and consulting services offerings. Oracle recognizes that every implementation is a multivendor implementation and is focused on the single outcome of a successful customer launch and adapts to the needs and desires of its customers from engagement to engagement.
As CSPs seek to serve more verticals with industrial IoT and private networks, how are SLAs and monitoring/service assurance going to have to change to serve the needs of those customers? Who needs visibility, into whose infrastructure and/or software, in order to successfully deliver and maintain private 5G networks?
To best understand how SLAs and monitoring/service assurance must evolve as CSPs serve more verticals with industrial IoT and private networks, carriers must look to the consumer digital domain. This might seem counterintuitive, but one could argue that the service user/consumer ultimately already has very distinct service, device, and software expectations as a result of our current technology climate.
Consequently, carriers must entirely rethink the access, ease, and flexibility extended to customers to be successful. The expectations are already set. Customers want to have full sight of the offering and options, the supply chain delivery, and ultimately the quality and outcome delivered. They already get this in their daily retail experiences: curated research and buying, on-time delivery with detailed tracking from supplier through shipment to delivery on location, and finally activation, personalization, and ongoing, intelligent in-life management. Service assurance will be the cornerstone of automated operations and closed loop automation in 5G networks for carriers. Best-of-breed assurance solution vendors like Oracle ensure service providers have a dynamic view of inventory, network automation for self-healing, auto-scaling, and closed-loop control of services. This enables them to intelligently re-configure customer and network services throughout their lifecycle, guaranteeing SLA compliance while reducing potential incidents of disruption or downtime.
How would you categorize network automation capabilities for monitoring/optimization at this point in 5G? How real and/or capable are AI and ML of taking over from human network operations teams? Do you think deeper adoption of network automation and/or predictive analytics will be driven by maturation/increased capability of the underlying algorithms, or are they already there and it’s human hesitance to rely on them that is the bigger barrier?
Today, automation capabilities across the 5G ecosystem are in very early stages, with monitoring and optimization still in infancy. AI and ML are entirely capable of taking over certain activities from human network operations teams, but the network is not yet ready to realize this promise. There are three primary hurdles that must still be overcome to accomplish this: technology enablement, solutions implementation, and human change.
Technical feasibility is still most limited by the practical reality that few (if any) services are rendered by sufficiently modernized network and systems solutions that deliver full automation, programmatic exposure, and open visibility. That’s not to say it isn’t a highly desired technology or that incredible progress hasn’t been made in network automation today. OMDIA and Oracle Communications found in their recent survey that 5G core automation will play a critical role in telco transformation. In fact, 92% of operators surveyed agree that 5G core automation is integral to any successful automation strategy. Cloud native will be integral for speeding the delivery of digital products with scale, resiliency, security, and agility. A 5G core network enables cloud services to be easily integrated and used in different services. This leads to more frequent and higher quality software releases, as well as full test automation and operations, something Oracle customers are already leveraging in their 5G network rollouts.
Looking for more insights on optimizing and monitoring 5G networks? Join the RCR Wireless News webinar on March 15th featuring AT&T and Viavi Solutions.
