By processing massive data sets using AI and machine learning, CSPs gain a high-resolution image of the network, and the insights that accompany this image allow them to pivot and adapt the network to meet the growing demands of customers and enterprises. The result is improved reliability, security, and importantly, the customer experience. This data-driven approach is key to a trusted operating model that will better position CSPs to meet the future needs of their customers.
The 5G future of operations will be automated – and why CSPs need to plan for it now
5G delivers an entirely new class of powerful technological capabilities and innovative business models. However, these advances come at the price of exponentially greater complexity and scale and require a new approach to automation to capitalize on the potential of 5G and offer new services. The CSPs that take the time to reinvent their operations environment now will reap the benefits of streamlined networks and differentiated services in the future.
5G, cloud and programmable networksare transforming the way CSPs build and operate networks and services. Whenthese networks are layered on top with modern business and operations systems,CSPs will finally have a powerful telco cloud and application stack that trulyenables them to offer on-demand network services. For example, Vodafone isalready working on their Networkas a Platform vision, and T-Mobilehas launched IoT Developer Kit, leveraging Network as Platform forInnovation. As a platform company CSPs can offer network services on-demandenabling enterprises to consume network services much in the same way theinfrastructure, platform and software applications are offered and consumed asa service and paid for on a consumption basis. Most critically, the success ofthis transition can only be achieved with a modern automated orchestration andoperations solution, and charging and billing capabilities to monetize thoseservices.
CSPsare in the midst of a significant journey to the cloud and all indicators pointto an acceleration of this journey in the coming years. However, CSPs arefacing an interesting conundrum when it comes to choosing the right type ofcloud - public, private or hybrid. The public cloud has been a very popularmodel for many IT workloads, and hybrid clouds and homegrown private cloudshave addressed a niche set of requirements. In 2023, CSPs will increasinglyconsider the private cloud option (like the Oracle Cloud@Customer) for OSS/BSS workloads as they try tobalance the need to achieve the same high level of agility, performance andscale, autonomous operations and reduced infrastructure TCO of a public cloudwhile addressing the pressing requirements of data sovereignty, residency andconnectivity concerns.
Digital twins have been employed inother industries such as automotiveand manufacturing to great effect. CSPs have asemblance of a network digital twin in its inventory systems, but the sheernumber of systems, siloed data, poor data quality and frequency of updates havegiven them a bad reputation. However, with the emergence of technologies suchas network telemetry, advanced data analytics based on ML and AI, and cloudbased data platforms, there is now a real opportunity to create a more unifiedand accurate, near real-time network and service topology of the dynamicnetwork and the services that ride on top of those networks. With thiscapability, CSPs will be able to create a true network digital twin, which willsupercharge a whole plethora of use cases such as intelligent networkinvestments, efficient network design and planning, advanced service impactanalysis and automated operations.
As we head into 2020, it is clear that CSPs are racing to deploy 5G networks. But the transition is complex and costly. (Read my most recent blog on 5G for greater detail on this topic.) In addition to bearing the high cost of acquiring spectrum, CSPs need to invest heavily in core and radio access networks, transmission, and infrastructure. 5G typically requires 5 to 10 times more cell tower infrastructure, and some industry analysts predict that network-related capital expenditures will increase significantly, roughly doubling the cost of ownership in some markets. How each CSP rolls out their 5G plan will be crucial.
To adopt a zero-touch service model, operators would move aggressively toward simplified product- and service-agreement portfolios, supported by fully automated, AI-enabled, cloud-based processes (Exhibit 4). This digital-first strategy departs from the familiar approach of incrementally making manual work more efficient through piecemeal offshoring and outsourcing. Operators such as BT and MASMOVIL have embraced a radical, future-back redesign of their service operations, enabling them and others to drive down costs (up to 35 percent in one case) while improving cycle times, improving accuracy, reducing call volume (up to 50 percent in another case), and increasing NPS by 20 points.
Our survey of more than a dozen global carriers currently planning their edge strategies suggests that most telcos will likely work with a partner to launch edge-based services, since they lack the technical depth needed to do so on their own. One of the most attractive options is teaming up with a hyperscaler, such as Amazon Web Services (AWS), Microsoft, or Google, that has the size and existing infrastructure to make edge work. But these tech giants could also end up becoming rivals.
As also noted in this same report, many of the individual elements of NLA are being worked on by TM Forum members, including: AI-based automated network planning; flow-through network build; automated network configuration/build/turn-up; automated end-to-end network testing/validation; AI operations predictive and preventive operations; AI-based, cross-domain correlation of events; zero-touch provisioning; self-healing and self-optimising networks; cognitive smart field operations; unified network and IT operations; and continuous monitoring and improvement of automation through agile delivery processes and tools.
Service and network assurance modernization is crucial for many CSP objectives such as automated operations, improvement in customer experience, intent-driven orchestration and enhanced 5G services. CSP CIOs can use this Market Guide in planning and sourcing assurance solutions.
Assuring that those small cells and backhaul links will meet service level agreements at turn-up presents a new challenge. Traditional manual provisioning and testing methods are simply impractical for current and future needs.
Imagine a visit to the doctor. You wait an hour for an examination at one location. Additional tests are needed at a different department. So, they give you a piece of paper to take to a different location, but the instructions were unclear. So, you return to the starting point to try again. Healthcare systems are modernizing to eliminate these layered inconveniences and improve their operations; with routed optical networking, communication service providers (CSPs) are optimizing in a similar way. The modernization of the Internet will require the most advanced silicon, optics, systems, and software to automate its operations.
MF: First of all, virtualised environments will be a lot more complex because there will be no standalone boxes and services will instead be delivered on top of shared resources. Secondly, one of the biggest causes of service problems is change. The ultimate goal is for CSPs to become more agile and in the future changes will occur much more frequently, making it even more imperative to use automated, active testing to ensure that services are delivered right. Without testing after a change has been made many customers could be affected before the problems are detected.
Several telcos are already working with IBM to implement their 5G networks. Chief among them is Verizon Telecommunications Inc., which announced at MWC that it has selected IBM and its subsidiary Red Hat Inc. to build and deploy an open, hybrid cloud platform with automated operations and service orchestration that will serve as the foundation of its 5G core. 2ff7e9595c
Comments