Parts of GFS GFS (Google File System) is a reliable distributed file system that is built on top of unreliable commodity machines. Commodity machines break down, and they break down pretty often, and given that there are many such machines, at any given moment it is a miracle if all the machines are working fine. So, it must constantly monitor itself and detect, tolerate, and recover from component failures regularly....
Map-Reduce Before getting into MapReduce, let’s first understand what distributed systems are — and why they’re widely used. Imagine You have some work that can be done by a computer. What do you do? Easy — you just assign the work to your computer. Now the work gets harder, and it’s taking longer to complete. What do you do? Simple — buy a better CPU, GPU, more RAM. Upgrade your computer....
Sorting and Aggregates Sorting What is sorting? Sorting is just arranging elements in a specific order. But why is this important for database management systems (DBMS)? Why does sorting matter in DBMS? DBMS needs sorting for two main reasons: SQL uses the ORDER BY clause to return data in a specific order. Sorting helps make query processing more efficient. When records are sorted, operations like joins and ORDER BY can be done faster....
This blog summaries the lecture 8 of cmu - Intro to Database Systems along with some things i learned on the way. Concurrency and Parallelism Parallelism: Parallelism means running multiple tasks simultaneously. It can be achieved with the help of multiple cores, multiple threads, GPUs, etc. Example: In programming, running two different functions at the same time. Concurrency: Concurrency refers to the ability to run tasks out of order without changing the final outcome....
Indexes in DBMS Indexes are very similar to the index section of a book. In books, indexes tell us where a word has been used and provide the page numbers. Just like that, indexes in a Database Management System (DBMS) tell us on what page or block a particular record is present. So why do we need indexes? Indexes save us a lot of time. Instead of searching from the start to the end of the book for a specific word, we can simply look up the index and go to the specific pages....
This blog is a continuation of the database systems series, covering Lectures 5 and 6 of the Database Systems CMU 15-445/645 course. Buffer Pool The buffer pool is the memory present in the main memory (RAM) that is used to store pages that are most frequently used, similar to a cache. It is managed by the DBMS (buffer pool manager), because the DBMS can do a better job of managing pages than the OS....
This blog is a continuation of the database systems series, covering Lectures 3 and 4 of the Database Systems CMU 15-445/645 course. How does a DBMS store data? A DBMS uses disks for data storage and persistence, and there are many storage devices that the database can use to store data. Storage Devices Storage devices are devices that are used to store data. There are mainly two types of storage devices: volatile and non-volatile....
I am currently studying database systems. This blog summarises lectures 1 and 2, as well as the SQL homework from the Database Systems CMU 15-445/645 course. It’s fascinating to realize the extent to which we interact with database systems in our daily lives. They’re not just confined to large-scale applications but are also embedded in our browsers and mobile devices. Before the advent of database management systems, data management was complex, with a strong coupling between the data and its management....
TCP stands for Transmission Control Protocol. It’s built upon the IP layer and provides features like reliable packet transmission, flow control, congestion control, and congestion avoidance. Why is TCP the most used protocol? TCP is widely used because of the features it offers, such as reliable packet transmission, flow control, and others. Reliable transmission ensures data arrives complete and in order, making it crucial for applications like file transfers, email, and web browsing....
How information is sent across the internet? Internet is a network of networks, billions of computers interconnect make the internet, information is sent across the internet in the form of packets, packets can be thought of a little containers of data, if you want to share an image to your friend, first the large image is broken down into number of packets, and these packets are transferred to your friend’s computer, in your friends computer they are reordered and assembled together(not necessarily) to form the image...