“Lucy in the Sky with Diamonds,” a song written by Lennon–McCartney (Beatles’ 1967 album “Sgt. Pepper’s Lonely Hearts Club Band”) and HAL 9000 Artificial Intelligence?  What do they have in common? Not much you may say. Well, let’s consider the context of fiction.

The Beatles song was based on “Alice in Wonderland”, a fictional novel and a fairytale, written in 1865 by English mathematician Charles Lutwidge Dodgson under the pseudonym Lewis Carroll. Alice in Wonderland tells the story of what happens when you fall through a rabbit hole into a fantasy world populated by peculiar creatures.  Artificial Intelligence was very much fiction up until a few years ago, and often tied to HAL 9000 from the movie “2001: A Space Odyssey” (1968), thus fiction or fairytale.  Well, fast forward almost 50 years. Alice in Wonderland is still fiction, but AI (HAL) has very much become a reality within networks.

Recent technology advances have made it more important for networks to handle a massive uptake in interference, traffic, capacity and connections. This is especially true for Wi-Fi networks. Qualcomm recently announced the new 802.11ax Wi-Fi SoC portfolio (IPQ8074) for network infrastructure devices (sampling mid 2017). The new 802.11ax portfolio promises to deliver 4x more capacity, and 4x faster connections.

As we consider the continued and massive need for bandwidth and spectrum, 4x more capacity, and 4x faster connections sounds great. Well, it is.  You need this to handle increased usage of HD video streaming, video-conferences usage, content sharing, social media and file transfers, etc. But, 4x more of everything also means more work and headaches for IT people. Why, you may ask?

The next protocol is 802.11ax with support for 12-streams (eight 5GHz and four 2.4GHz), 8×8 MU-MIMO, 80 MHz channels and other features to maximize capacity and coverage.  The protocol aims to improve traffic scheduling with a managed approach of resource allocations. In addition, 802.11ax will include optimization of wake-up time which can possibly reduce Wi-Fi power consumption by 2/3rd to extend device battery life.

AX promises device data rates of 1.8 Gbps peak speeds through Dual Band Simultaneous (DBS) combining 2.4 and 5 GHz bands, and higher order modulation (1024 QAM) and access point data rates up to 4.8Gbps.

Ax in the Sky with AI
As Wi-Fi and small cell licensed and neutral host networks become more crowded, dense and diverse — a myriad of devices with differing protocols (802.11 a,b,g, n, ac… and soon ax) with data rates ranging from 2Mb to 1.8Gb, and spectrum bandwidth ranging from 1×1 40 MHz to soon, 12×12 configuration with eight 80 MHz streams, we’ll need AI more than ever before. After all, old and new devices must be able to connect and work. The world would be silly simple if there was just one protocol. But, that’s not the case, and we need cloud and AI functions to help make it all work, even with available Wi-Fi and small cell SON features.

AI Powered Networks Becoming a Reality
AI can improve overall network experience when we consider that 75% of network issues (source: Gartner) are identified by the end users. How? AI can proactively manage and fix network issues. AI enables big network thinkers to rethink what is today, and improve resources and people more effectively to compete and deliver better products and services on a global scale. In the coming 5 years, Artificial Intelligence (AI) has the potential to completely change the way we run, deploy, manage and utilize networks, and ramp up new global services.

KodaCloud’s system is AI powered and automates time-consuming troubleshooting and optimization tasks. AI makes decisions regarding which parameters to tune based on patterns fed by the learning module. The system also creates client behavior profiles, and uses fingerprinting to tailor device-to-AP interactions. AI learns network behaviors and patterns by gathering data from KodaCloud’s global network.

In summary:

  • Rule-based algorithm in the cloud
  • AI model that updates the parameters of the rule-based algorithm
  • RRM enforcement, measurement and data collection modules on the AP (enforcing and collecting data)
  • Automation of all functions without any knobs to control, or conflicting manual parameters to tune

All networks are managed by one global expert system for proactive network monitoring and troubleshooting using patented Artificial Intelligence (AI) — which automatically corrects 80-90% of network and connectivity issues. The system collects detailed information from each Access Point, and networks and presents the information to the AI system that reacts at machine speed, capacity and cost. With the forthcoming 802.11ax protocol and neutral host small cells, AI is increasingly important for enterprise, hospitality and service provider networks to manage the massive increase in performance, traffic, capacity and interference.

All companies, large or small, would benefit from including AI in network and services strategies. AI can be the foundation for growth and productivity for years to come.   “Lucy in the Sky with Diamonds” is still a great Beatles song. But, the AI-powered network is no longer a fairytale.

Ronny Haraldsvik
Twitter @haraldsvik


Chat with a Wi-Fi ExpertAsk us a question