Strengthening next generation network with Artificial Intelligence and Machine Learning

Strengthening next generation network with Artificial Intelligence and Machine Learning

Last month (April 2018), I represented Aricent at the MPLS+SDN+NFV World in Paris and in attendance were some of the world’s largest mobile operators including AT&T, DT, Turk Telecom and Reliance Jio. SDN/NFV is coming to fruition and 5G is almost here. The main theme at the show was automation - and what transpired was that while many operators have been making good ground, end-to-end automation of the network was a conundrum.

The network operators we spoke to at the congress in Paris, share the same concerns as their peers across the world and understand the challenge on the horizon. It is the rising levels of network traffic and the complexity that it brings. As handsets grow in sophistication, subscribers have grown an insatiable appetite for data. According to Cisco, mobile data traffic is expected to reach a whopping 49 exabytes per month and exceed half a zettabyte annually. The zettabyte era is almost upon us and operators will have to manage an increasingly complex network that has been flooded with a cocktail of data traffic.

At MPLS+SDN+NFV World, Aricent highlighted a technology that provides light at the end of the tunnel for network operators. Network AI helps operators to create a brain for the network, using advanced machine learning algorithms, so that the network can self-manage and self-protect by analyzing petabytes of network data and taking intelligent decision real-time.

A smarter, stronger WAN

Aricent takes a comprehensive, forward looking approach to R&D and what we found in our labs is that by integrating Artificial Intelligence and Machine Learning, we strengthened the SD-WAN.  Automating SD-WAN with self-learning capabilities, thanks to AI and ML, enables zero-touch network operation with intelligent feedback loops.

During our session in Paris we shared our experiences of building an autonomous-WAN for a leading telco. Here are the top 5 benefits our autonomous-WAN solution delivered:

1. Enhanced end-user experience and reduced customer churn: It proactively detected network issues and dynamically met subscribers needs

2. Achieved higher operational efficiency: Our solution simplified network operations and delivered a consistent, error-free network

3. The assurance of a guaranteed Service Level Agreements (SLA): An autonomous WAN solution can negate the financial impact of SLA penalties and improve the brand reputation

4. Boost employee productivity: Automating the network frees up expert resources from day-to-day tasks like debugging and root-cause-analysis

5. Gain a competitive advantage: Operators can rely on trusted and proven telecom expertise to have a robust, secure and future-proof network in place that is head and shoulders above the rest

Enhanced security for the network

Another concern voiced by operators attending the Paris event – and shared by carriers across the world – is security. Distributed Denial-of-Service (DDoS) attacks are all too common and operators have to constantly monitor for unusual traffic patterns. It is sometimes not easy to distinguish a spike from a DDoS attack and say, high demand for a popular album from Taylor Swift or Drake!

Thankfully ML algorithms can interpret vast amounts of traffic behavior and an autonomous-WAN can predict performance issues before users are affected. In this way, connections with algorithms that scrape Twitter feeds will confirm the hypothesis. Have hacking groups been threatening action against a particular enterprise? Or are millions of excited “Beliebers” clamoring for the Justin Bieber album in the weeks leading up to the spike? The autonomous-WAN will analyze and adapt accordingly, either shutting down ports to isolate the DDoS attacks or adding bandwidth to accommodate the surge from subscribers.

NetAnticipate

At the show, Aricent also presented the NetAnticipate Framework. NetAnticipate is a highly scalable, intent-based, predictive AI platform for processing petabytes of network data in real-time for accurately predicting network anomaly and take preventive measures through automated feedback loop, thus realizing  self-learning network. NetAnticipate supports use cases from variety of domains, ranging from, optical transport networks, 5G, IoT, datacenter, security and many more. The framework has been designed to seamlessly manage the growing influx of data traffic. The carrier-grade solution enables end-to-end automation by combining AI models, semantic telemetry and intent-based orchestration capabilities. NetAnticipate improves efficiency and enhances the customer experience with faster troubleshooting and reduces the time it takes to resolve problems

The future is here

Forward looking service providers such as AT&T and Telefonica are already preparing for the zettabyte era. It will be impossible for humans to analyze the colossal volumes of data that will traverse the networks and make an informed decision, in real-time, within seconds. Traditional tools today do not have the power to analyze heterogeneous data at ultra-high speeds and mobile operators can ill-afford network downtime or poor quality of experience. At best, frustrated subscribers might sound off on social media. At worst they churn in droves.

The time has arrived to strengthen the core network with automation to self-configure, self-manage, self-heal and self-protect itself with zero human intervention. Automating the network is not a ‘nice to have’ option. It is a pressing necessity.

About the Author

Subhankar

Subhankar Pal
Director - Technology, Innovation
Aricent

Subhankar Pal is Director Technology in Aricent, global design and engineering company, and has 18 years of professional experience in IT & Telecommunication industry. In Aricent he is part of CTO organization. His responsibilities include providing technology thought leadership and incubation of innovative solutions, in the area of SDN, NFV, machine learning & artificial intelligence for telecommunication industry. His area of interest includes advanced machine learning in 5G, SDN, NFV and IoT networks, for real time & predictive analytics. Subhankar has extensive experience in speaking at international conferences and presenting technical papers in various forums. Some of his recent talks were at MPLS+SDN+NFV World Congress in Paris in April 2018 and CORD Build event in Sant Jose in November 2017. Subhankar has whitepapers published in IEEE forum & other international events in the areas of cloud, IoT & unified communication.

 

Leave a Reply

Your email address will not be published. Required fields are marked *