 Hi, my name is Sam Thompson, and at Asia Crypt, I'll be presenting the work Practical Proably Secure Flooding for Blockchains. This is joint work with Chenda, Liuzang, Christian Matt, Uli Maurer and Guillermo Rito. To motivate this work, I would like to look at the example of blockchains. So a blockchain protocol is a protocol where a group of parties interact in order to build and extend a total order. However, it's not always guaranteed that this total order will be built. This is only guaranteed if each party owns some resource, and a majority of these resources remains under honest control. But this is not the only assumption needed for blockchain protocols to work. In fact, they also rely on the flooding network. So in order to build the total order, both of these assumptions will need to hold. So let's have a closer look at this flooding network assumption. So flooding network is a very simple protocol where any party is allowed to input a message and as a consequence, all other parties will receive this message within some delta time. And this is the assumption that is used in order to prove the security of blockchains in a wide range of work. Alright, so let's look at how such flooding protocols are actually implemented in practice. If P1 wishes to send a message, then they will choose a random subset of the parties as their neighbor and forward the message only to these neighbors. This will continue on, such as now when P2, P3 and P4 has received the messages. Well, then they will forward the message to a random subset of the parties again. This continues on until all parties have received the message. It's not difficult to see that if a large fraction of the parties are actually dishonest, then either P1 has to send to really many parties or otherwise the protocol will not work. In this case, P5 and P7 will not receive the message. With this in mind, let's revisit our blockchain example from before. We had the blockchain all to be the total order relied on the majority of honest resources and on such flooding network. Now, if we instantiate such flooding network using the traditional approach, then this in fact implies that we need an assumption on the number of honest parties and not only on the resources. So it's naturally wondered that we could build a flooding protocol that will in turn build a total order only based on the majority of the honest resources behaving honestly. So in this work, we asked the question, can efficient flooding be realized assuming a constant traction of honest weight? The answer is sounding yes. So our contributions are that we first provide the new flooding protocol called weighted fan out flooding. This has the properties that it's secure assuming any constant traction of resources being honest. It has logarithmic diameter, which means that delivers messages to all parties very fast. And it has message complexity that is comparable to state of the art that relies on a majority on some assumption on the number of honest parties. Additionally, we provide extensive simulations of the weighted fan out flooding protocol and this confirms the practicality of our protocol. Thank you for your attention.