mySQL, NoSQL, and now NewSQL?

by Eric C
In today’s fast paced world with data growing at an exponential rate, a database must be scalable and perform well with today’s demands in storing data. SQL databases have been around for decades and the basic architecture wasn’t created with scalability in mind. Apparently there have been new advances in database technology and it includes instances of NoSQL and NewSQL. Michael Stonebraker is a seasoned database creator who is now a chief technology officer for VoltDB, explained the benefits of using “NewSQL” to better benefit today’s demands for database performance. Stonebraker stated that traditional SQL systems have many limitations and that includes performance. SQL databases are also not scalable onto more than one server. If one were to make an SQL database scale onto more servers for better performance, it would be very complicated to manage. NoSQL was indeed created to improve on scalability and is increasing in popularity; it also has its own limitations as well. The main problem with NoSQL is that it cannot perform complicated mathematical queries. However with NewSQL, it improves on all of the issues with SQL and NewSQL, making it a more efficient database system that can process requests faster and can scale to more than one server. According to Joab Jackson, the author of the article entitled “’NewSQL’ Could Combine the Best of SQL and NoSQL” from PCWorld, using NewSQL “can execute transactions 45 times faster than a typical relational database system” and “can scale across 39 servers, and handle up to 1.6 million transactions per second across 300 CPU cores” (Jackson).

This article from PCWorld is related to this week’s topic mainly due to type of database management system used for databases. Since we are learning about the relational database management system, I found it interesting to know that this type of database is decades old technology. Although there are some disadvantages, it is still the most popular DBMS used today. It is amazing to believe that such an old database system is still currently in use today with little changes.

As I discovered in the article, database systems this old wasn’t created with scalability in mind and have performance issues in today’s world of growing data. As a result, traditional DBMS cannot be used for all situations, which is why there are variants that improves on the old.

Sources:

Jackson, J. (2011, August 24). ‘NewSQL’ could combine the best of SQL and NoSQL. Retrieved October 7, 2012, from http://www.pcworld.com/article/238728/article.html

Menegaz, G. (2012, October 1). What is NoSQL, and why do you need it?. Retrieved October 7, 2012, from http://www.zdnet.com/what-is-nosql-and-why-do-you-need-it-7000004989/

4 thoughts on “mySQL, NoSQL, and now NewSQL?

  • October 7, 2012 at 8:14 pm
    Permalink

    Interesting article. When I was looking for an article this week I was reading one about noSQL and it being able to be used across different servers while still keeping the data on all of the servers relatively uniform in structure. It seems like newSQL could be a good way to increase efficiency in databases by spreading the work load over so many different processors.

  • October 7, 2012 at 11:49 pm
    Permalink

    Very Interesting. The traditional approach of SQL databases means companies will have to purchase more servers as their traffic increases which can be problematic because it will most likely interfere with the day to day task of running the company. NoSQL solves the problem by distributing data among different servers as data increases. They both have their pros and cons. One company might prefer a wider network over a handful of servers. Another company that carries high-risk information might want to store their data on their own private servers.

  • October 8, 2012 at 1:05 am
    Permalink

    As new problems rise, technology will evolve and to solve or provide efficiency. It’s great to know that database technologies are advancing to accommodate the vast amount of data. However, new scalability would require better algorithms.

  • October 8, 2012 at 4:53 pm
    Permalink

    Modern RDBMS systems can also distribute data among different (logical and physical) servers and they usually do. We will learn more about it in Chapter 5 Physical Database Design and Performance. And each RDBMS release has always been added new features and enhancements to address the need to store more and more data. For example, Oracle8i was released in 1999 and could handle databases with terabytes (1024 gigabytes) of data. In 2001, Oracle9i was released and could deal with up to 500 petabytes (1024 terabytes). Oracle Database 10g offered support for exabyte (1024 petabytes) databases. Oracle 11g now offers more. Most business won’t even come across too many databases with exabytes of data right now, and vendors like Oracle and Microsoft will keep supporting them. And for larger amount of data they use to forecast or analyze for trend or decision making we will use data warehousing and mining techniques which we will talk about in Chapter 9. While it makes sense to use NoSQL in some situations (1. User-generated content causes headaches and issues of scaling for “read-heavy” websites; 2. business that has significant amount of data change overtime; and 3. companies like Google and Amazon have to use NoSQL to meet their needs of scale), 90%+ of the database in use today are still RDBMS. Some says that Non-relational database is really not a replacement but rather a supplement to RDBMS, i.e., a hybrid ecosystem. So don’t worry about you are learning an obsolete technology in class, you are really just getting more and wider knowledge about database.

Comments are closed.