MySQL Room for Improvement

Gaston Nusimovich
Software Architect at a tech consulting company with 51-200 employees
Only one of the engines, InnoDB, supports ACID transaction control. The best performance engine, MyISAM, has no transactional control support. It would make a lot of sense to include one engine that offers both very good overall performance and transactional control support. All versions of MySQL, including the Community Server version include in its design as a Pluggable Storage Engine Architecture. This architecture allows for the support of multiple options of Storage Engines, so, a software architect can design a sofware solution with MySQL based on a specific Storage Engine that is capable to cater to a particular type of storage use case requirements, like for instance, web applications with heavy read workloads and moderate write workloads, or intranet desktop applications with heavy write workloads. Depending on the version of Community Server installed is the list of available Storage Engines: All it takes is the execution of the command SHOW ENGINES to retrieve the list of installed Storage Engine plugins. When you create any given table in MySQL, you have the option to select what Storage Engine (among the engines available) will be applied to said table. If this option is no explicitly mentioned at the end of the "CREATE TABLE" command, the current default storage engine will be assumed. For web apps with intensive read workloads and moderate write workloads, the engine that offers a good overall performance is the MyISAM engine, but this does not support ACID transaction control, and it uses a table-level locking mechanism (thus, from a performance standpoint, this is not proper for apps with heavy write workloads). The InnoDB engine uses a record-level locking mechanism, so, it is proper for apps with heavy write workloads. It also supports Isolation Levels, which is important for applications that have many clients doing read and write operations concurrently. The other storage engines offer support for very specific use cases, like for instance, flat file tables (CSV engine) or memory-only tables (Memory engine). Getting back to the InnoDB engine, it offers features that are similar to the database engine in SQL Server, and according to some of benchmarks that we have run, SQL Server Express provides overall better performance than MySQL Community Server with InnoDB tables. My "Room for Improvement" comment is that it would be great if in the future Oracle were to provide with the Community Server edition a version of InnoDB with better overall performance, while still requiring a small hardware footprint. View full review »
Muhammad Nurazhan Moin
Senior Web Manager at a university with 501-1,000 employees
It would be helpful if there were a graphical user interface to administer, configure, and tune it. If it had something similar to Microsoft’s DTS engine then it would be the best database system out there. View full review »
Partner at a tech services company with 51-200 employees
* I think a better front end would be a better solution (web application front end, similar to what Red Hat is doing to Fedora). * Another nice solution for MySQL clustering would be the use of Webmin. * Also, security measures could always be improved, and the clustering process could be enhanced as well. I recommend using UFW, iptables, and firewalld. View full review »
Find out what your peers are saying about Oracle, Firebird, PostgreSQL and others in Open Source Databases. Updated: November 2019.
377,029 professionals have used our research since 2012.
Lucas Holt
Senior Application Programmer/Analyst, Team Lead at a university with 10,001+ employees
The weakest link is replication. MySQL’s replication is touchy and doesn’t support master-to-master setups. MySQL’s replication was implemented by playing back SQL statements getting executed on the master database node. That means if a slave node gets behind, it has to catch up by running every SQL query that has executed on the master. When the slave gets too far out of date, it is unable to catch up because the replication log is lost. Master to master replication is used to allow writes to either node in the cluster (INSERT, UPDATE, DELETE). Since MySQL does not support this, you can only safely send SELECT queries to slave nodes. If you run an insert, update or delete against the slave node, then it is out of sync with the master and may break if a later replication event comes from the master on the same table. View full review »
QA Lead at a logistics company with 1,001-5,000 employees
I feel that some tools which make it easier to create queries or make it easier for other functions would be really interesting to see. View full review »
Tomas Dalebjörk
Information Technology Technician at a tech vendor with 51-200 employees
How to scale out with shard clusters. The way how to handle replications needs to be improved, as we noticed that there are some performance drop of (using only one core for example) View full review »
Zeger Knops
Head of Business Technology at a tech services company with 501-1,000 employees
It does not stand out regarding scalability. When the company size increases, the user base having actual experience with (very) large MySQL solutions is reduced. View full review »
Abhishek Ghosh
Director at RC Ghosh Group of Companies
MySQL needs improvements on its diagnostic features. View full review »
Miguel Araújo
IT Manager
Manzeel Uprety
Co-Founder at a non-tech company with 1-10 employees
The only service which could be improved is its usability. The entire user experience needs to be revamped to meet the 2018 design standards. View full review »
Haim Tzadok
Co-Founder/CEO at Grigale LTD
SQL and NoSQL use. View full review »
Yongho Gil
User at a software R&D company with 10,001+ employees
When working with a cluster wide, I have to use the MySQL cluster version. View full review »
Find out what your peers are saying about Oracle, Firebird, PostgreSQL and others in Open Source Databases. Updated: November 2019.
377,029 professionals have used our research since 2012.
Sign Up with Email