At Digital Transformation Middle East this month, Anuradha Udunuwara, Senior Engineer / Solutions Support at Sri Lanka Telecom will be discussing network transformation and automation. Read some of his insights on the topic below before the big event, and register today to see him and many more of your industry peers share crucial insights.For some of you, 5G will have been home for Christmas last December. But for some, 5G seems to be few Christmasses away – for many reasons. 5G after all, is not about a technology, speed, but a brand new user experience. Though there is an involvement of the technology, mainly the 5G NR (New Radio) and 5GC (5G Core), we easily forget the transport (or
x haul) network in between.
As we envision a future where we have remote control of machines in mining sites, we’re performing a real surgery remotely, controlling autonomous vehicles and many more over a last mile 5G wireless connection, the ability to have all these different applications running across a network with their respective quality of service (QoS) parameters is actually achieved through network slicing. The required capabilities like virtualization, automation, agility and flexibility is achieved through network
softwarization (SDN, NFV and Cloud).
Beyond 5G for network transformation
By the way, 5G is not the only use case for network transformation and softwarization. As operators transform themselves from communications service providers (CSPs) to future digital service providers (DSPs), they realize that they need both agility and flexibility which is only possible in software. Hence the softwarization of telco networks and systems.
Network transformation cannot happen overnight. Whether we like it or not, we will always have to live in a hybrid transition/migration period. Our challenge will be to shorten this period. From today’s hardware world of physical network functions (PNFs) and hardware defined networks (HDNs) to tomorrow’s fully software and cloud world of software defined networking (SDN), virtual network function (VNF), cloud-native network function (CNFs), we get a transition period with a mix of both worlds.
The HDN/PNF world could be complex, but we mastered that complexity. SDN/NFV looks complex, mainly because of the lack of knowledge, skills and education. What would be the complexity associated with a hybrid world? How do you make sure your network functions (NFs), both that of 5G related or not, residing in physical, virtual and cloud environments live in a perfect orchestration so that the end-to-end service chains are possible and visible, in this disaggregated functions mess?
It’s not easy, given that your PNF, VNF or CNF could be anywhere in a bare metal, VM or container, with many possibilities of implementation options.
Further, there are many open source options to do the same thing. We are yet to see the consolidation of efforts in the open source community. The best example is
ONAP and
OSM.
The vendor perspective
Next comes the vendors. Given, they also hold their own business challenges, their true willingness to be open is still a question. As the margins of selling hardware boxes and software are getting thinner, mainly due to open hardware and software initiatives, they now slowly move to the service business (training, system integration, operation and maintenance of telecom networks and systems). Anyway, in world where everything is about time, TMT (Time to Market) is the key for telecom products. This requires capabilities like automation, DevOps and agility and achieving those in a multi-vendor environment is challenging.
As you can now see the real problem for network transformation is now becoming multi- dimensional: multi-environment, multi-vendor, multi-technology, multi-standard/open source, …. Anyway, this is the reality and we have to face it, if we are to transform ourselves. Even though the environment is hybrid, we have to implement automation for onboarding, especially VNFs and CNFs, service chaining and networking etc. If you are really particular about TMT, leveraging onboarding automation of NFs is key. But this requires new skills, knowledge and education. In the absence of these new capabilities within the CSP environment, vendors capitalize the opportunity and that sometimes deprives the CSP ever achieving the required capabilities.
On one hand CSPs want to release products faster to the market to start generating new revenue. On the other, the internal capabilities, mainly human resource is stopping them doing that. The ideal plan should be to start with a help of a reliable vendor, but have a two to three year plan to build the required capabilities in house. If the CSP continues to depend on a third party, then they will not be able to sustain in the long run. Of course you need vendors to supply you the hardware and software. Do you want them to do your day today work as well? Then how do you define your business model?
Automated onboarding
One of the essential requirement to have is a NF onboarding automation tool that supports both VNFs and CNFs (or 2 separate systems). But you need to check whether they are really open, meaning that they can onboard any open VNF/CNF and optimize the process/operation. Further, it should be able to perform automated service [chain] design as well. Ultimately the entire lifecycle of a NF from procurement/sourcing, onboarding, testing, acceptance/validation, deployment, configuration and assurance/maintenance should be automated, leading the industry to a same-day NF onboarding status.
OPNFV Verification Program (OVP) is a helpful program in this regard.
Network transformation with lifecycle automation
This has to also be coupled with end-to-end dynamic service orchestration. Two notable initiatives in terms of standardizing and supporting this industry requirement are the works of
MEF and TM Forum. MEF has developed
LSO (Lifecycle Services Orchestration) specifications with open APIs to automate the entire lifecycle for services orchestrated across multiple provider networks and multiple technology domains within a provider network and across provider networks. TM Forum has developed the
TM Forum Open API program. We see the industry is adopting these common language frameworks.
Closed-loop service assurance
The optimum delivery of services in reality is also linked with closed-loop service assurance. AI can be used to achieve this easily. As operators bring real automation to their service engines, service assurance data will play a key role.
Ideally, service assurance data will have to be in a “closed loop" that uses constant assessment of network health (traffic and performance) to optimize service parameters like quality, detect possible threats and do many other tasks automatically. This plays an important role in service providers' ability to offer new services which change with customer needs, without requiring network reconfiguration or a lot of human touch. For example, Sri Lanka Telecom was a champion in the
Proactive Service Assurance via Closed Loop Predictive AI/ML TM Forum Catalyst project. The project demonstrates advancement toward fully automated service assurance for network transformation, working with actual network data and utilizing AI, predictive analysis, and knowledge-based self-learning.
As we move from yesterday to tomorrow, the hybrid infrastructure in a telco environment will pose new challenges in automation, especially the onboarding of network functions, achieving real-time dynamic orchestration of end-to-end service and closed-loop service assurance. However, with open collaboration among all stakeholders of the industry, we see a silver line that makes us navigate the transitioning period for network transformation slowly and steadily. Tomorrow will surely be brighter than today.
