If we know the time it takes to deliver any message in the network, we also know the time it takes for a block created somewhere to be delivered anywhere else in the network. When using Bitcoin, it is recommended to wait for k equals five blocks to be appended before a transaction can be considered committed using the figure that we explained in the previous model. We have to wait for a transaction to be included in a block of the chain and for five subsequent blocks to be appended after this block. The overall expected time is one hour. This works pretty well as long as after one hour, we're guaranteed we will know of all the conflicting blocks at the index of the transaction. The assumption made here is called communication synchrony. As we covered earlier in the course, in communication synchrony, we require two things. First, that every message sent gets delivered in a maximum amount of time, and second, that this maximum amount of time is known by the algorithm. However, by now, it's clear that block chains are typically used for the internet network. So, is it really realistic to assume communication synchrony on the internet? Today, we'll cover this topic. Sometimes, natural disasters have dramatic consequences on the time it takes to deliver a message over the internet. As an example, in 2017, typhoons damaged an underwater cable located 54 kilometers away from Hong Kong. This cable was used to transmit internet data between Sydney and Hong Kong. This outage impacted the quality of the connectivity of Australian internet users for several weeks. Sometimes, humans misconfigure the routes that convey the information between internet computers. As an example, in 2008, Pakistan telecom wanted to ban a trailer on YouTube. The company decided to redirect the traffic from Pakistan internet service providers so that users could not watch the content. It turned out that their redirection was announced, by mistake, to a large internet node. This node then started redirecting the YouTube traffic to the same root, therefore, impacting the connectivity of two-thirds of the internet population to YouTube, stopping the communication for two hours in Asia. If a similar misconfiguration were to affect Bitcoin, then two hours delay, would be more than enough to wrongly believe that the transaction is committed, while other conflicting blocks may remain unknown. These examples are representative of scenarios that happen every year and disrupt the communication time on the internet. There are eight well-known fallacies or wrong ideas that people tend to have about distributed systems and networking. These happen when people ignore the way they are implemented. These common fallacies include the idea that the network is reliable, latency is zero, bandwidth is infinite, the network is secure, topology doesn't change. There is only one administrator, transport cost is zero, and that the network is homogeneous. So let's correct some of these wrong ideas. Firstly, it's important to note that the network is neither secure nor reliable. This means that communication messages can be lost or intercepted. Sending a message takes time and has associated costs. Bandwidth is also limited. This means that sending large blocks typically takes more time than sending small blocks. Additionally, there is no single administrator, and different parts of the network offer different characteristics, which makes their configuration error prone, as we have just discussed. These fallacies are the assumptions that overlying application should not make in order to prevent problems. To conclude, communication synchrony is very helpful. It allows us to define how to detect that a transaction is committed and that the assets it transfers are successfully transferred. Although communications synchrony is helpful, it is not realistic. There are too many factors that affect the delay of messages. Whether these are expected congestion, natural disasters, human misconfiguration, or malicious activities. All these are quite frequent. In the next lecture, we will discuss solutions to this problem.
