IoT Lab 10

Hashing is a topic that's relevant for many areas of computer science

• Hash Tables: By hashing values and storing them in a table, we can make the average value of a lookup go to O(1) simply because hash functions are computed on the CPU in a very small amount of time, and requires ideally just a single memory access to verify the correct element is found and return.

• Corruption: In general human interactions, the method of confirming if a piece of data (ie a telephone number) was correctly transmitted, is by reading the data back entirely. This works somewhat acceptably for a short transmission. An example of where this fails is when one pronounces a number consistently, but poorly, leading to the same data being copied incorrectly in both transmissions. For longer pieces of information, such as the installer for a piece of software, this can reach into the gigabytes of data, which can be intensive to download once, let along multiple times. Instead, we compute a hash of the data, by applying a mathematical function, and compare the results of the function on both the originating system and the receiving system.

• This is also commonly seen with the hash-art one sees when generating SSH keys and when attempting to log into systems via SSH. By allowing humans to verify that the SSH key art appears normal, one can reduce the chances that they are in a MITM attack.

• Password Storage & Comparison: It is widely known that password transmission and storage in its original form is a generally poor idea. Instead, we transmit hashes of the salted password, making correlating passwords/hashes very difficult, yet being able to quickly validate user credentials.

• Bitcoin: The bitcoin paper is a very interesting read and is not too technical.

Although I did not have the necessary hardware to run the second half of the lab, it was interesting to read about Iota and how comprehensive the libraries are for block chain and secure hashing.