Daniel Karp, Partner, Cervin: So first things first - Tell us a little bit about yourself and about Lightlytics.
Or Shoshani, Co-founder & CEO, Lightlytics: My name is Or. I’m the Co-founder and CEO here at Lightlytics. Briefly about myself - I used to serve in the Israel NSA department and was an engineering team leader with a branch that implemented routers and switches with security algorithms. Later on, after 5 years serving in the Israel NSA, I decided to run my own company. My previous company was bootstrapped, and we started providing services for the Israeli Homeland security agencies and to the Israel air force division, providing them with dedicated security algorithms. After 3 years running as a bootstrapped company, we decided to pivot ourselves into a product lead company and we implemented networking capabilities in hardware adding fundamental elements in intercommunication that are in data centers today. We began partnering with Mellanox, and as the relationship grew and things went well, Mellanox gave us an acquisition offer and we accepted it. After the acquisition, I joined Mellanox as part of the business development group. I used to work with large scale EOMs there like IBM, Dell, and Nvidia (this was prior to Nvidia’s acquisition of Mellanox). We were building huge data centers, super-computers leveraging the networking capabilities of Mellanox, the compute power of IBM and the GPU’s of Nvidia. During one of the deals that we unfortunately lost to Cisco, I met Stav and we instantly clicked. Stav is now serving as CTO and co-founder and the relationship began there. We also collaborated with Liran who is now serving as the VP of engineering and is the 3rd co-founder of Lightlytics. We started Lightlytics 2 years ago with the vision to empower the cloud operations teams to easily deploy to and operate the cloud. We tackle one of the biggest challenges in today’s cloud operations which is the growing complexity of the cloud. Lightlytics is building a simulated model, in real time, of the entire cloud architecture and by building that we help them throughout the entire Ops flow by planning configuration changes faster and troubleshooting faster, We’re building impact analysis across vulnerability, security, compliance, cost, capacity and even performance divisions.
DK: There’s a lot to unpack there. One of the things that you mentioned is that at Mellanox you ran into a set of problems, but also ran into your co-founder and you decided to basically leave Mellanox, in order to tackle this pain point. Can you talk about what inspired you to actually undertake that pain point and why didn’t you do it with Mellanox?
OS: This is a great question, Daniel. So I think it was like winter outside and raining outside, and we had a phone call from a huge customer in China, experiencing downtime in production, and after a week of hair pulling, they didn’t know what to do. They didn’t know what the change that caused the downtime was, so after a week, they just decided to shift the blame to Mellanox, and told us that we had some issues in our switches. So we decided to take a flight to China and help them debug their data centers. Stav was literally inside the data center debugging, and they were so desperate that they decided to shut down half of the data center to understand where the issue was. Eventually, after like two weeks of debugging, they found out it was misconfigurations in one of the switches that caused the entire data center to go down. We were asking ourselves - How does this multi-billion dollar company with thousands of engineers have this type of issue? Our team, at the time, was also working with Microsoft Azure, Facebook, and Alibaba and we were exposed to how they deploy and maintain their infrastructure. We noticed that there is a huge gap between the hyperscalers and the enterprise market, and we decided based on that to tackle this gap and help the industry tackle issues with these configurations, whether to predict misconfigurations or to do impact analysis to ship faster and to close the loop.