Compare And Contrast The Performance Of Different SQL/NOSQL Solutions On Different Workloads
Z1DONUK3
Proposal.docx1. Motivation for solving the problem.
2. Method using which the problem will be solved.
3. Datasets that will be used to evaluate the solution. Remember to argue why your data is sufficient, i.e., representative of a broad sample of data over which the solution you proposed will be applied in real life and not an anecdote.
4. Metrics that will be used to evaluate the efficacy of the solution.
5. Experiments that you will run.
6. Progress you have made in the last month.
7. Schedule listing milestones for your progress.
Compare and contrast the performance of different
SQL/NoSQL solutions on different workloads (e.g., MySQL
vs Postgres vs Hive vs Pig vs Voldemort vs Mongo vs ...)
1. Motivation for solving the problem: The rational sql database has been used for a long time and it is suitable for most of the database. But in recent years, NoSQL has been used more and more often and SQL fall short. We aim to figure out the differences between these two types of database by doing experiments in this project. When we’re dealing with specific problems in the future, as data scientists, we can always choose a proper design.
0. Method using which the problem will be solved: We will try different workloads in different features with either SQL or NoSQL databases. The datasets will vary from large to small, simple to complex, and they can vary from frequent manipulations. We would compare their efficiency in different queries using both databases. And then we can draw conclusion by comparing efficiency results with corresponding databases.
0. We decided the dataset will be used to evaluate the solution will be company’s product with customers’ First name and Last name.The reason why we think the dataset will be sufficient is because there are numbers,dates, and groups of words and numbers called string, structured data could be displayed in titled columns and rows that were easy to order when we use the company’s dataset.
0. We will test out efficiency on one computer with all the same environment. We will pick out several common queries that will most possibly use on the chosen dataset. And record the time the queries take and compare them.
1. Experiments we will test the time to write the generated key-value pairs per each database.It should be noted there is a significant difference between SQL and NoSQL. As the size of the database increase or decrease,the performance difference between database become evident.
2. In the last month we got to know what is a nosql database and why the huge companies choose to design their own nosql database instead of using sql database. We also learned index and hashing used in relational database, OLAP and OLTP, as well as distribution.
3. Milestones: We check our progress every week, starting from 20/2/2018 that the proposal is due. We will also get the dataset test and analyze the report by 03/16/2018.
