Mobile Network Automation
with Artificial Intelligence

Voted the #1 AI Startup by Mobile Operators and Investors
Selected as One of the Worlds Most Promising Growth-Stage Companies
@ GSMA MWCA 2018


MArket Size: $50B (CAGR 47.6%)
Founder with Domain Expertise who felt the pain
Successful POC with Tier-1 converted into
7-figure commercial deployment
Self-learning AI with access to Petabytes of data oN Tier-1 Networks


The Founder

Miro Salem, launched the world’s first LTE networks, and operated the world’s biggest networks as an engineer, subject matter expert, and technical advisor to the industry’s top executives.

+ The Founder's Story (click to open)

For many years I was a wireless engineer. I still am. I worked for the worlds top vendors and carriers, like Ericsson, Verizon and AT&T.

I did every job related to wireless engineering and mobile network management and operations.

I'm an immigrant who came to the US to study Electrical Engineering, and in my last semester I fell in love with wireless protocol. There were two competing technologies, WiMax and LTE. I liked LTE more, I liked the mobility part of the protocol. I am a US Citizen now, but when I graduated I was on a F1 visa (student visa) and it wasn't easy to get an H1-B (employment visa).

I started my career as a drive-tester for Sprint. What is drive-testing? Remember that Verizon commercial: “Can you Hear me now”? It’s like that but with a car and a cell phone connected to a laptop. I drove every other street in Omaha – 10 hour days over 9 months. It’s a notoriously painful and tedious job. But I didn't do that for long, 9-months later I landed a job at Nortel and then Ericsson. I started as an RF (Radio Frequency) Engineer 1, and after some intermediary jobs (see below) I went on to launch the world’s first LTE networks. I launched LTE for Verizon, AT&T and then following the launch I was asked to stay and do post-launch optimization.

I became a Subject Matter Expert (SME) after I realized that I was good at recognizing patterns from one carriers faults and then recognizing them again on another carrier and advising the other carrier to prevent the problem from happening in the first place. So I became a subject matter expert, and technical advisor to the world’s largest mobile network operators.

Throughout my career we used antiquated software, shell scripts, spreadsheets, and VBA scripts, all at the same time. Our screens we're open with tab after tab of spreadsheets, leading to our common nickname - spreadsheet engineers.

Until one day, when I looked around, and saw how we were working, small rooms, elbow to elbow, late at night, red bulls next to us, manually, and reactively trying to solve network problems. Problems that impact thousands, if not millions of people. That's when I knew something had to change. And I couldn’t change it from within.

I founded 3TEN8 with one goal in mind, a self-operating wireless network. To develop an AI, that self-learns from network data, and proactively addresses network degradations. An intelligent system that can predict network outages, before a single customer is impacted, or a single negative tweet is sent. An AI that can benefit from collective intelligence, to learn from one network and apply what it learned to another.

As a result, an often stagnant, and slow-moving industry will be fundamentally changed and ready for the future use cases of 5G and beyond.

Some of the jobs I worked in network operations:

  • Drive Testing

  • NOC Alarm troubleshooting

  • Network Design & Capacity Planning

  • Launched First LTE in the World (Verizon, AT&T)

  • RAN Black Magic!

  • Pre-Launch Tuning

  • Post-Launch Optimization

  • Interference Hunting

  • Single Handedly Reconfigured every site in the United States (Verizon & AT&T)

  • SME LTE for Verizon and AT&T

  • Advisor to Verizon and AT&T Executives and Engineering team

  • Carrier Grade Wifi for Time Warner Cable

New business models, including network slicing, will support new markets and impose unprecedented operational agility and higher cooperation across network domains. The resulting exponential increase in overall complexity makes automation a necessity.
— ETSI

Purpose

Our mission is to fully automate wireless network operations using artificial intelligence. We’re creating a Self-Operating AI for mobile network operations.

+ more

We believe that carriers in the future will outsource the management and operations of their networks to best AI in the market. It’s an economic principle that its time is coming. We, today, are creating the best self-learning and self-operating AI in the market ahead of this rising need. Read on.

+ Problem

The way we operate mobile networks today is very inefficient and ineffective. Very manual and using rule-based systems. This inefficiency is driving the cost of operations so high that, despite making $1 Trillion in revenue, carriers are struggling to make margins. We are reaching a point where network complexity is exceeding humans ability to manage mobile networks.

This increased in complexity and inefficiency is causing engineering teams, at carriers, to be in a constant state of fire-fighting, only getting to work on a problem after it has already impacted customers. It’s like being in a state of postmortem analysis. This is tarnishing carriers brands and further adding to their losses.

The current solution, is to throw more bodies at the problems. Since industry vendors that provide equipment and services to carriers make 70% of their income from selling human services, they have no interest in introducing efficiencies that would cannibalizing their own revenue.

Compounding to the problem, the introduction of more complexity in mobile networks. More parameters can be configured and more metrics from more network elements that are being measured. This surge in data and configurable variables are overwhelming the engineering teams that are operating these networks. 5G deployments (i.e., IoT and machine type communications), are projected to add billions of new devices to these networks, increasing the complexity of managing networks even more.

5G network slicing turns the traditional one network owned by one carrier into many slices that are leasable. Businesses will be able to lease a slice of AT&T’s network in the future and specify the quality of service that they need. That means that now engineers will have to manage many slices of networks instead of the traditional one network (one slice), and therefore adding even more complexity to network management.

Needless to say, this is a lot of data to be considred before making operational and management decisions, and although efforts have been made to store this data in Big Data solutions, little has been done to innovate operations.

network-complexity-figure.png

Solution

Using AI to automate network operations.

AI Automation for Network Ops

3TEN8 AI learns from internal network sensor data as well as external environmental data to automate network operations.

 

Traction

  • 3TEN8 successfully completed a paid POC with a Tier-1 carrier

  • The POC has converted into a nationwide commercial agreement

  • 5 additional POCs have now been added to 2018/2019 with projected conversion in 2019/2020

  • Currently generating revenue and projecting 7 to 8 figures in 2019

Market Size

The Autonomous Mobile Network Operations market size is a $150 Billion in 10 years. The market is growing at a rate of 47.7% annually. 3TEN8 will also consider future deployments in WiFi (MSO), and Enterprise I.T.

Market Size

Autonomous Network Operations is a new and growing market

CAGR 47.7% (MarketsandMarkets)

Product

  1. Self-Healing

  2. Self-Optimizing

  3. Self-Configuration

All towards the ultimate destination of Self-Operating networks.

Product Milestones achieved:

Self-Healing – Our last POC with a tier-1 operator 3TEN8 AI was able to predict 70% to 90% of faults and outages and delivered actionable insights to prevent incidents from happening. This was provided with a average of 40 hours of lead time (early warning). This milestone is significant as it is the first in the industry to provide predictive maintenance and operations which is a significant step towards self-operating AI.

Business Model

Pricing is per device managed per year. All sales are at minimum 7-figures.

Devices include – Subscribers, IoT, Autonomous Vehicles

As of April 2018, there are over 8.5 Billion connected devices in the world and 5 Billion unique subscribers. (https://www.gsmaintelligence.com/)

Team

Miro Salem - Founder & CEO (Ericsson, AT&T, Verizon)

Dr. Elica Rahdar, Ph.D. - Co-founder & COO (UCLA Neuroscience, Research & Stats)

Naresh Guntupalli - Co-founder & VPE/ML (ex-Paypal)

+ 3 Data Scientists (Stanford, Netflix, Facebook)

Investors

Alchemist Accelerator – 500 Startups – Citrix – Amr Awadallah (CTO Cloudera) – S+C

Resources

http://www.snstelecom.com/network-complexity-5g-rollouts-will-drive-son-self-organizing-network-spending-to-5-5-billion-says-sns-telecom-it