Innovation Conference: Ofer Bengal, Founder and Chairman at Redis.

Thanks for having me, glad to be here.

Well, you know, it's hard to say the nature of people, but I would say that from a very early stage I was designing wooden cars, wooden models of cars and things like this, and I did it with cardboard and some other things, and obviously wood. So the roots were there, but other than that, it's just something that comes out of you natively, I would say.

So yes, at the age of 18 I decided to go to the academic reserves here in Israel. So I went to the Technion, studied aerospace engineering. Ask me why aerospace engineering? I have no idea, I just selected that. And I was in the Israeli Air Force for like nine years, engaged with designing and developing airplane modifications, particularly around avionics, weapon systems, etc. And as part of that, before Israel got the first F-16, I was sent for a long period to General Dynamics, the company who basically made the F-16, designed and manufactured the F-16, to learn firsthand from the engineers who develop the airplane all the insides and internals of the airplane, so we could later on make modifications to the airplane. And as a matter of fact, when I came back, we improved and upgraded the airplane, especially around the weapon carriage system and things like that. So yeah, this was my start. And after the Air Force, I basically decided to carry on in this space and started my own company, which did more or less the same thing. So this was a consulting company that did work for defense companies in Israel and in the United States, and basically did projects around airplane improvements and so on.

Yeah, so you know, everything is coincidence in life. I was traveling at that time; we had some defense companies as customers of this consulting company of mine, and I was spending a lot of time in the US. And in the morning, you watch TV before you go to meetings, and you see ads on television primarily for toys because this is the time for moms and little kids, etc. So I was intrigued by that. The richness and variety of toys and ideas and concepts and everything went hand in hand with TV series and etc. So this really intrigued me, and I started to think about toy ideas, and I became some sort of toy inventor. So I invented a few concepts for toys and was successful in selling them to some of the leading toy makers in the US. So the first one was basically something like miniature laser tag, if you remember laser tag, but before laser tag. So the idea was to put backpacks with a battery on the back of action figures, and then from there there is a wire and a miniature gun which you put in the hand of the action figure. When you press a button on the back, it fires some sort of infrared beam, and if it hits another figure like that, the other figure does some sort of funny noise and is paralyzed. And then I expanded it to several miniature vehicles with guns like this, so that the boy can do some sort of mini war. At that time, all my mind was in wars because obviously I was born in Israel. But later on, I completely left that area and made some board games and puzzle games and things like this, which I sold to companies like Hasbro, Milton Bradley, and so on and so forth.

Yeah, so as I said, everything in life is coincidence. And dealing with toys, one thing that frustrated me was that I was sort of an inventor, whereas I truly wanted to be a manufacturer. I wanted to do the business around it, etc. So I decided to start a toy company out of Israel that will not just design the toys but also manufacture them, market them, and sell them globally. And I wrote a very nice business plan and went to raise money. At that time in Israel, there was only one venture capital firm, no one remembers that one. So I went to them, and at that time no one knew the term 'angels,' which is so popular today. So it was just rich people; we just went to rich people to look for money. So I went to a few of those, and one of them was Zohar Zisapel, who started, him and his brother are considered the forefathers of the Israeli tech industry. They started in the 80s a tech company called RAD Data Communications and were very successful in that, made a lot of money also. So I went to them to look for funds for my new company. So I told the story and brought the toys to show them, etc. And Zohar said, 'You know what, give me the toys for the weekend. I'll give it to my kids, and if they like it, I will invest.' That was the due diligence, if you will. So he took the toys, he came back the next week, he said the kids love the toys, let's invest. And then a few days later, he calls me and said, 'You know what, I have an even better idea. How about starting together a data communications company? I will put the money, and you will be the founder and CEO.' So I said, 'You know, it doesn't make any sense. I have no idea about data communications, and actually at that time I didn't like computers so much even.' So he said, 'You'll learn.' Initially I said thank you, I'm not interested, I would like to pursue the toy thing. But after I struggled to raise enough funds to start a toy company, I sort of gave up, and in a way unwillingly took his offer and started a data communications company by the name of RAD Technologies. And that company dealt with management systems for the infrastructure of large networks, like very large local area networks and telephone networks, etc. And we were pretty successful for that time. In 97, I took the company public at NASDAQ. According to today's standards, it's funny: our AR at that time was $40 million, we raised like $25 million, which today you raise in a seed round or in a round, but we raised it in an IPO. But it was fun, it was good, and I managed that company for 10 years and then left in the year 2000 to do other things.

Sure. So after, in the year 2000, I was managing RAD Technologies for 10 years, and then this was the time when the internet exploded in the world. And at that time, there were starting to talk about the internet moving to mobile, to the mobile phone. Now this was obviously six or seven years before the first iPhone. At that time, there was a protocol which was called WAP. And I decided to leave RAD Technologies and to start another company around mobile internet. So that company was called Mobile Economy. We raised seed funding and a round, etc. We built nice technology around that, a WAP gateway which was installed in many of Europe's mobile operators. But we struggled to generate revenues because no end users were using mobile internet at the time, pre-iPhone. So in 2003, I fire-sold that company to one of our partners, brought some money back to the investors, tagged myself first time as a failure, and moved on. And then at that time, I was invited to be on the board of several tech companies in Israel and also served as some sort of fractional CEO for a few companies on behalf of certain venture capital firms in Israel. And one of those companies which I was on the board of was called Crescendo Networks, and it was started by a guy by the name of Yiftach Shulman, who eventually became my partner and co-founder in Redis. So he was the co-founder, I was a board member in Crescendo, we became friends, and we decided that one day we will do something together. It took us, I don't know, like three or four years until we really did. But in 2011, we started Redis. And the beginning was kind of unusual. When we started, the idea was basically to accelerate applications. And when we analyzed the delay and what happens in the end-to-end response time of an application, we realized that the major delay is in the data layer. And the data layer or the database industry at that time was dominated by Oracle, and it didn't make any sense at that time to start a database company because it was considered insane. So we started to accelerate applications by using a caching system which was very popular at the time, which was called Memcached. And this was another open source way before Redis. And we decided to take that and basically make it more robust and more suitable for enterprise use. So this is how we started the company. The company at the beginning was named Garantia Data, sort of like 'we will guarantee your data' or something stupid like that. But anyway, after a short while, we discovered Redis. Redis, the open source, was started by an Italian guy by the name of Salvatore Sanfilippo in 2009, basically two years before we started our company. And it took time until it got market traction, but we identified it at a very early stage and we loved it, and we decided to take it as the engine under the hood of our Memcached product. So basically we were selling something with a Memcached API, but underneath we used Redis to power it. But after a while, we saw that open-source Redis is becoming very, very popular, so we decided to basically change and become a company initially called Redis Labs, and later on we changed the company name to Redis.

So the open source model started with very good intentions, which is to enable the masses to enjoy the work of individuals, and that was very good in principles until the cloud era. Now in the cloud era, and it started with AWS, but Azure and Google Cloud basically did the same: they were picking up successful open-source products or projects, commercialized them, and started to sell them to their customers. Now because of their dominance in the market, it was very difficult for the people who basically started the open source and built companies around them to compete with the hyperscalers because of their dominance. So in the case of Redis, for example, just to mention here that a few years after we started the company, the creator of Redis, Salvatore Sanfilippo, joined our company, became part of our company. So it was kind of weird: he was leading in the company a group of developers who continued to develop the open source, and at the same time, whatever we put out there in the open source was immediately adopted by the hyperscalers. So we felt stupid: we are working hard, they don't contribute anything to the open source, they take 95% of the commercial market whereas we take 5%. It didn't make any sense. So we started at a very early stage, and by the way, this was not just us. We talked to some other companies in the space, companies like MongoDB, HashiCorp, Elastic, and the like, and they all suffered from the same thing. I think that we were the first that came to the conclusion that it doesn't work and we need to do something about it. And we came up with another license which is a source-available license, meaning that anyone can get the source code, but it comes with certain limitations. The only limitation basically was that you cannot take the product as is, commercialize it, and resell it. However, if you have a certain app, I don't know, a travel app, and you want to use the product even commercially, you can do that. So anyone basically can do whatever they want with this license except taking the product as is and reselling it. And we were followed by companies like MongoDB and Elastic, each with their own flavor of the license, so it's not the same license basically, but all of them were around the same principle, which is to avoid the hyperscalers using what we produce. Now practically or formally, this is not open source anymore because according to the OSI, which is the Open Source Institute, which is basically the one who is carving the guidelines and everything else around open source, if you add any restriction to open source, this is not open source anymore. So basically today, Redis and all the others are not open source anymore. But in practicality, for a regular developer or even a company, it doesn't make any sense because they can take this source-available code and use it without any limitation.

So what we saw that these hyperscalers did not change anything fundamentally with the product. They just wrapped it with some sort of service layer and offered it to their customers. And we thought that the open source as is is not good enough for enterprise use. If you talk about things like high availability, scalability, governance, security, and so on and so forth, we came to the conclusion that heavy enterprises will find it difficult to use the open source or even the products offered by the hyperscalers. So basically we redesigned the infrastructure layer of Redis. We designed a brand new clustering technology for the product, high availability technology for the product, and some other things. And we targeted from day one the niche of the most demanding enterprises in the world, which are not happy with the sort of 'good enough' product provided by the hyperscalers. And if you look today at our customers, I would say that over 50% of Fortune 100 companies are customers of ours with Redis Enterprise, and actually more than 50% of Fortune 10 companies, companies like Apple and the like, are also customers of ours. So this was our strategy to compete with the hyperscalers.

So the beginning was not so rosy. It's very funny because now we are such good friends, but at the beginning, we met Salvatore in one of the trade shows and just made acquaintance. And when we started to use Redis under the hood for our product, I don't think that we even told him. We simply started to use it. And then after we saw that Redis is becoming very popular, we decided to change the name of the company to Redis Labs. And at that time, we did not even tell Salvatore anything about that. We simply went ahead and changed the company name to Redis Labs. A day later, we get an email from him saying that he's not accepting this, and if we are not changing the company name from Redis to something else, he is going to change the name of the open-source project from Redis to something else. This was very funny. So I took a decision to revert, go back to our original name, and wait until we start a discussion with Salvatore. And I think that we met Salvatore face to face, and then he was convinced that this is a serious business, we are not just a bunch of kids trying to steal his work or something, but contributing a lot. We contributed a lot. I mean, after Salvatore himself, we were the second largest contributor to the open source. And then we gave him a few options for the name of the company, and he accepted Redis Labs. And not just that, a couple of years later, he joined the company, and a couple of years later, he transitioned all his rights, trademark rights, and copyrights to the open source, etc., to the company. So yeah.

Yeah, and this touches upon a very important point. I mean, whenever you have in parallel an open source project and a commercial product, the requirements are never aligned. Because Salvatore and his team talk to the community and they get requests from the community, whereas we talk to very large customers and get requests from them, and they are not the same requests obviously. The requests from the enterprise sector are around governance, security, that sort of thing, whereas the community is interested in other things. So yeah, there was always a conflict between the two, and we had to somehow navigate our efforts and resources between the two. No one was fully happy, but that's life.

That was easy from day one. So Redis originally, the open source, started as a caching layer, basically accelerating other databases. What we have done over the years, as part of the commercial product especially, was to make Redis a standalone database, a source of truth. Now the use cases for Redis were always around performance, like latency and speed, latency and throughput. So Redis was always the fastest database on Earth. And with Redis Enterprise, we were able to process even hundreds of millions of transactions per second with an average latency of less than one millisecond, and there is no other database in the world that can do that. So the use cases are not necessarily conflicting with those of companies like MongoDB and others. I'll give you an example: when you go to a point of sale and give your credit card or Apple Pay, and when you do a transaction and the transaction needs to be authorized by the credit card company, Visa and the like, what happens behind the scenes is that they need to look at all your history with them and all the history of the business, and take all this information, process it with their engine, and make a determination whether to authorize the transaction or not. Now all that needs to happen in a split of a second, otherwise no one has the patience to wait. The only database that can do that at scale is Redis. And basically all major credit card companies are customers of ours. So this is something which is less suitable for databases like MongoDB and others, but they have many other use cases where they are more suitable than Redis, because the problem with Redis is that everything runs on DRAM, in memory, and memory obviously is more expensive than SSD or disk, especially when you're talking about very large workloads. So there are many use cases where it makes much more sense to use other databases which run on other media. I would say here that in recent years, we also ported Redis to run effectively on SSD, almost the same performance as on DRAM, at a much lower cost. But still, it's different use cases. So I don't think we see ourselves competing head-to-head on the same deal. This happens very, very seldom, to compete in the same deal with MongoDB, for example.

Yeah, absolutely. And we have started to work on this, I would say prior to the ChatGPT bang. We always thought that Redis can be very powerful as a vector database, so we made Redis a vector database at an early stage. Today, it's definitely the highest performance vector database. So capabilities like vector database, RAG, vector similarity search, feature store, semantic caching, etc., are things that we do with Redis today. By the way, OpenAI is a large customer of ours. So as you can imagine, many prompts by people in the world are repetitive. So instead of running the whole big model over and over again on the same prompt, they cache those prompts and responses, and this is called semantic caching. And basically, some of the responses that you get from ChatGPT are from that cache rather than the model itself.

First of all, be prepared for a bumpy road. A startup company is always challenging. You have days where you are in euphoria, you have days where you are in depression. So you need to forget all that, stick to what you do. Don't be deterred by competition, because whatever you do, whatever you start, there will be other companies chasing the same field, the same technology. Just be determined and do what you have to do. The other thing is to start looking at the GTM at a very early stage, because most people who start tech companies come from a tech background, and many of them, at least first-time founders, lack experience with go-to-market, and how to know what is the exact market that they target and how to approach it, and so on and so forth. So these are things that need to be addressed at a very early stage. And especially today, what I find is that there are many startups which I would call 'on steroids.' They start a company, and like three years after, they are at $100 million in AR, which was never heard of in my time, like 10 years ago, for example. And I think that this is attributed primarily to the fact that these founders were experienced enough and started to build the GTM function at a very, very early stage, way before the product was ready for the market. So when the product is there, they can go full force and start to sell.

I would say that this guy, Zohar Zisapel, that we mentioned before, was an inspiration for me. Because he started a tech company out of Israel at a time that most of Israel's exports were cotton and oranges and things like this. And the idea that you can basically invent or develop technological products in Israel and sell them globally was mind-blowing for me at the time. So yeah, unfortunately, this guy passed away a year ago. We remained good friends until the very end, him and his brother. They both passed away. They are still considered the forefathers of the tech industry in Israel. So yeah, he was definitely my mentor.

So two years ago, I stepped down as CEO of Redis after 12 years. I became the chairman, and we hired a very powerful CEO to lead the company to its next phase. His name is Ran Trope, and he was previously a CEO of a public company by the name of 59, and previously a high-level executive in some very large companies. And I became a chairman. A chairman is a very pleasant and good life basically; you don't do anything except when it comes for board meetings, you need to prepare a little bit, participate in the board meeting, and so on and so forth. But this was not enough for me. So in these two years, I traveled the world, I've been anywhere, I did a lot of fun, but I felt that something is really missing because probably in my nature I need the action. So two months ago, I started a new company around the intersection of enterprise data and AI. I think that there is a lot to do on that front. So I partnered with two other great guys, and we started this company. We raised the seed funding for this company, and these days we are actually starting to work on this new concept, which at least for now is revolutionary. We'll see how long. As I said before, you are never in a vacuum, and competition starts fairly quickly. But we will move fast and hopefully build another great company.

Yeah, so honestly speaking, I didn't have time recently to read many books. But there is a book that I like very much, and they made a movie out of it. It's called 'Being There.' Have you heard about this? It started as a short novel by a Polish writer by the name of Jerzy Kosiński, and then they made a movie out of it, which I highly recommend that you look for it. It's something from, I believe, the 70s or maybe the end of the 70s. But it's a really amazing book. It's about a simple-minded gardener that all he knows is how to water the flowers, etc., but he meets some people who listen to him and think that he's a philosopher. I wouldn't do a spoiler here, but eventually this guy almost becomes the president of the USA. So very recommended to see how the human mind works. Also very funny, with great actors. So recommended.

Thank you guys. Thank you. See you soon.