This Startup Invented a New AI Technology to Provide Complete Network Visibility for Smart Cities and Critical Infrastructure

Share on linkedin
Share on facebook
Share on twitter

A new forecast estimates there will be 41.6 billion connected IoT devices, generating 79.4 zettabytes of data in 2025. Today, more than 7 million IoT devices are powered up daily and the weaponization of IoT devices is proliferating rapidly as more devices are integrated into networks. Threats from botnets and DDoS for hire services are almost ubiquitous, with high competition and wages as low as $5 per hour. As a result, smart cities are increasingly on their defense by a variety of threats, including sophisticated cyber attacks on critical infrastructure, industrial control systems, connected devices and sensors, or privacy sensitive data.

Prevailing cyber security solutions can protect against malicious network attacks, but can only attain partial network visibility, at the cost of hundreds of monitoring appliances. The leading provider of network visibility solutions is NETSCOUT, with security appliances designed to protect individual network devices, costing up to $100,000 per monitoring device. Moreover, each appliance is limited by processing specifications to 40 gigabytes per second, which quickly become obsolete as connected networks are constantly evolving; in size and traffic rates, as well as in their architecture, migration to cloud and hybrid environments. As a result, CISOs of smart cities are forced to compromise on their network’s security visibility accordingly. 

Changing the paradigm is Israeli startup Cynamics, coming off the heels of a seed funding round and newly invented machine learning technology to holistically monitor network traffic without any appliance, and at scale. “Cynamics can monitor your entire network’s traffic with only a 1% sample of the network traffic in order to provide 100% network visibility and security for threat detection” said the startup’s CTO and technology inventor Dr. Aviv Yehezkel. Their new approach to network security and performance grants smart cities and governments unmatched network visibility and super-early prediction of attacks, long before they affect the network, at scale and cost effectively.

Cynamics’ co-founders (left to right): Eyal Elyashiv (CEO) and Dr. Aviv Yehezkel (CTO). Photo: Ofer Keidar

The startup was founded by Eyal Elyashiv and Dr. Aviv Yehezkel in early 2019 to coalesce the pair’s expertise in network security and artificial intelligence. Elyashiv, CEO, is a seasoned executive with vast experience in the US local government and municipality sector and was COO of public safety and security startup Carbyne 911. Yehezkel, CTO, is an alumnus of Israel’s military Naval Intelligence and unit 8200 as a research captain, and led AI research at NICE Systems. An expert in big data and machine learning, he earned his PhD in Computer Science from the Technion at the age 24, the youngest faculty ever.

“Legacy solutions require thousands of appliances, and only provide partial visibility at best” explained Yehezkel. “Typically, prevailing solutions require all traffic to be constantly forwarded to the network visibility appliance, which analyses the network packets for visibility inside the network and provides alerts in case of a threat. While this approach was well-suited for past networks that had a small amount of network devices and simple architecture, it is no longer relevant or feasible in today’s complex networks. A bottleneck arises in the ability to watch the entire network, which is limited to the number of visibility appliances. Legacy networks, public clouds, private clouds combine together to form one large network that cannot be monitored at all.”

Cynamics captures small network traffic samples using standardized sampling protocols from all network layers, including IoT and wireless, as well as dark fiber to obtain unlimited visibility, preventing network blind-spots. They developed a series of innovative algorithms and technologies for inferring complete network visibility from the small collected samples,  for continuous network learning and threat prediction. By summarizing the network state in every timestamp and autonomously learning the most important network fields to look at, the startup’s AI solution also identifies hidden patterns in the network traffic that precede an attack across a city’s global network. “We only sample small fractions of the packets, amounting to approximately 1% or even less, and this is sufficient for our technology to infer the complete picture on network traffic and alert on suspicious patterns” explained Yehezkel. “The samples enable us to cut off more than 10 seconds or more from the attack detection time that we can provide, compared to existing solutions that need to collect, process and analyse each and every packet all the time, where every additional second on detection timing can translate into millions of dollars.”

“The paradigm shift is from local incisive monitoring of your network to a global view of the entire interconnected global infrastructure” said Elyashiv. “We want to provide network visibility on a massive scale in cities, and that’s a huge change.” 

A summary of the network at each timestamp for several network fields. Their AI technology automatically learns what are the most important network fields for the customers’ specific network based on traffic samples collected. Other AI models developed by Cynamics analyse the traffic change over time and predict whether in the next time period, there will be an anomaly of threat that will surface. Graphic: Cynamics.

Cynamics’ patent pending artificial intelligence algorithms analyse traffic patterns for anomalies, triggering appropriate actions in a fully autonomous manner. The SaaS solution requires no appliances, network modifications, or deployment and installation processes. The solution can be seamlessly integrated into the network non-intrusively without the need for any device or agent, and their on-boarding process, one of the startup’s technological breakthroughs, is completed in a matter of seconds.

Graphic: Cynamics.

To date, Cynamics has raised millions of dollars in pre-seed funding, and the company has been evaluated at $10 million. Their advisory boards consists of domain experts Prof. Eli Upfal, world-renowned AI expert and professor at Brown University, and Dr. David Primor, former executive in the IDF’s unit 8200 and the Israel National Cyber Security Authority.

The startup recently completed a successful pilot with a major city in the US. Their technology is on display for private demonstrations at the AWS re:Invent conference on December 1st, 2019. “We believe we’re going to disrupt the Network Performance Monitoring and Diagnostics industry with our new approach and paradigm” said Elyahshiv.

With the rise of 5G technology and virtualization, we can expect a massive explosion in the number of network resources, edge compute devices and IoT, all of which can be weaponized. And in this increasingly connected technological landscape, every smart city service is as secure as its weakest link. With Cynamics’ solution, that notion may no longer be accurate. 

scroll down to read more