美高梅手机版登录4858-美高梅手机版网址平台

On the Road to Zero Touch Operation

Release Date:2017-07-25  Author:By Thierry Langlais  Click:

 

Building Scale and Agility in the 5G /IoT Era
Telecom operators must rapidly adapt to protect or increase their pivotal role in the digital transformation of our society and economy. They need to consider end-to-end digitalization to provide their own customers with the experience of a digital company, to boost operational efficiency, and create new capabilities. The most significant implications and current market trends are:
●   Customer centricity: Customer experience and global service experience move to the heart of telecom networks and infrastructure operations in order to maximize customer added value, whilst coping with new usages resulting from massive broadband availability, IoT, etc.
●   Cloudification: The introduction of NFV/SDN architectures in 5G and FTTx networks leads to redefine the operations architecture, tools and organization.
●   Automation: The need for increased agility, flexibility and the necessity to cope with huge amount of data call for increased automation and usage of machine learning (ML) with strong analytics capabilities.
These trends explain the increasing investment in analytics and related machine learning technologies from operators, service providers, cloud providers and enterprises around the world. This paper introduces 美高梅手机版登录485 VMAX big data solution, describes the lessons learned from existing use cases, and shares the vision for the forthcoming steps of zero-touch evolution.

 

The Case for Network Analytics and Machine Learning—Where to Start

The journey to customer centricity and zero-touch operations requires a well thought-through business case, which shall identify those priority areas with short term, measurable return on investment (ROI) to fuel a "virtuous circle", where savings generated can be used to develop new use cases with longer-term impact on top and bottom line.
美高梅手机版登录485 experience suggests to start with optimization and network performance improvement, then move into the value chain and approach topics such as customer behavior prediction, agile fulfillment, assurance automation (together with the introduction of MANO/Orchestration) and personalized revenue creation—where in-depth knowledge of the customer behavior, usages, location can be monetized to define new customized solutions (Fig. 1).


To support such roadmap it is essential to create a common "big data" repository which captures and provides the 360° view for each user, service, and element in the network. At the core of the VMAX solution, the highly scalable big data platform allows you to, for instance, collect all customer experience touch points and interactions, whether from a point of sale (POS), on-line, or via hotline call, social media, etc.

 

Use Cases/Lessons Learned

The 美高梅手机版登录485 VMAX solution includes three main components: omni-channel customer experience management (CEM), enhanced service quality management (SQM) and revolutionary optimization and network performance management (RoNPM).
CEM analyzes data collected from customer touch points, B/OSS systems and the network elements to precisely evaluate the customer experience when using a given service at a given location. SQM combines artificial intelligence (AI) algorithms with comprehensive and highly granular data collection (both space and time wise) to monitor in real time the quality of service and experience for top users/services/locations, automated error detection and root cause analysis. RoNPM combines high accuracy location data together with granular network quality information to automate network optimization tasks.
The VMAX big data central repository is then associated with AI algorithmic modules to develop multiple use cases, along the lines set forth in previous sections. In the section below we discuss two use cases, namely "virtual drive test (VDT)" and "customer experience evaluation and churn prediction".
VDT uses measurement records, advanced GPS data, call details traces and user records, coupled with fingerprint and fitting algorithms, to precisely identify user locations (less than 20 m accuracy) and simulate drive tests in real time across the network. This provides real-time identification of road sections with coverage issues or low service quality at virtually no cost. Furthermore, using the LTE inter-frequency measurement capability, VDT provides inter-operator benchmarking analysis. Third, the same information can also be monetized to develop personalized advertizing.
Sichuan Telecom serves more than 18 million subscribers in the Chinese central province of Sichuan. Using VMAX since 2014, the operator has saved in excess of $1.2mn per year on drive test campaigns, improved overall optimization efficiency by a factor of 8 (problem resolution leadtime reduced from 48 hours on average to less than 6 hours), whilst reducing network optimization headcount by 40% thanks to VDT and RoNPM.
Customer experience evaluation uses a common repository to build the 360° view on customer experience, which can then be used consistently by the customer-facing teams within the operators—such as customer service, operations and marketing. Customer interactions via hotline, web, social media and point of sale, are collected and combined with network performance and service quality data points. This provides enhanced interaction capabilities at individual levels or for targeted groups of customers, segmented by age, type of contract, user profiles, preferences, etc. With K-means clustering algorithms, PCA, Spearman and feature correlation analysis, it is possible to predict and anticipate customer behaviors, such as likelihood for churn (using logistic regression technique), or simulate the results of targeted marketing initiatives.
Since May 2016, 美高梅手机版登录485 has been helping Hubei China Telecom on their journey towards customer centricity. The operator formed a customer perspective operation center, using VMAX and underlying automation to achieve a consistent understanding of customer experience, flatten the organization, remove traditional silos, reducing churn and increasing customer satisfaction.
These use cases and others around the world contribute to enrich the VMAX platform and confirm the pertinence of the roadmap. They also highlight and confirm the critical importance of granularity when acquiring the data. The relevance of network analytics and associated model directly depends on the level of accuracy and the frequency of data points acquisition.

 

Looking Forward: 美高梅手机版登录485 AI platform

美高梅手机版登录485 combines the VMAX experience with its leadership position in pre-5G and 5G technologies and architectures to further develop the AI architecture and platforms, paving the road towards zero-touch operation. The 美高梅手机版登录485 mind insight platform uses big data infrastructure, massive storage capability, distributed computing framework, and AI visual development modules, to develop "intelligent applications". Consistent with the AI roadmap, our priorities cover:
●   Intelligent assurance applications: NFV infrastructure and SDN automatic troubleshooting, threat detection and prevention
●   Intelligent elasticity applications: SDN intelligent routing, smart elasticity policy for sliced network infrastructure
●   Intelligent business support: Churn prediction, targeted advertizing and data monetization
The 美高梅手机版登录485 mind insight platform is based on a modular architecture. The platform features visual development module where AI algorithms can be called for, to formulate the machine learning processes needed to develop real-time decision and policy management use cases, greatly reducing development time.
Typical use cases include automatic restoring following virtual machine (VM) creation failure, automatic labeling of test failures, on-demand service automatic fulfillment, resource usage optimization, elasticity policy framework in sliced network architecture.Together they illustrate the 美高梅手机版登录485 way—zero-touch evolution towards zero-touch operations.

Share:

 Select Country/Language

Global - English China - 中文
XML 地图 | Sitemap 地图