The data management landscape is complex and quickly evolving. Nothing emphasizes this evolution more than the emergence of ‘super apps’ — or applications that process millions of user interactions per second. Factor in Big Data and the cloud, and it becomes clear that e-commerce merchants need a new generation of databases that can perform better and scale faster.
Any online business without an updated database is likely running MySQL, a database barely updated since its inception in 1995. After all, the term “NewSQL” didn’t become a part of the digital lexicon until Matt Aslett, an analyst for the 451 Group, coined it in 2011.
While MySQL is certainly capable of handling a good deal of traffic, as a business continues to grow, its database will probably reach maximum capacity and its website will stop functioning properly. If you’re unsure of whether or not your organization is ready for a NewSQL database, here are five signs you may be outgrowing MySQL:
- Difficulty handling reads, writes and updates – MySQL has capacity limitations. As more and more customers complete transactions on your website, it’s only a matter of time before your database stalls. Moreover, as your load increases, and you are finding it difficult to handle additional reads and writes, you may need a different database. MySQL can scale reads via “read-slaves”, but applications have to be aware that reads aren’t asynchronous with the write-master. For example, when a customer updates products in his or her e-commerce cart, it should be read from the write-master. If not, you risk available-to-promise quantities being wrong. If that happens, you’ll have a bottleneck in the worst possible place: your e-commerce checkout line. A bottleneck at checkout can result in abandoned carts, or worse, you’ll sell inventory you don’t have, and have to deal with upset customers, and possibly negative social media exposure.
- Slow analytics and reporting – MySQL databases don’t provide any real-time analytics capabilities, nor do they provide support for other SQL constructs. To address this problem, both Multi-Version Concurrency Control (MVCC) and Massively Parallel Processing (MPP) are required for processing massive workloads because they allow writes and analytics to happen without interference, and use multiple nodes and multiple cores per node to make analytic queries go faster.
- Frequent downtime – MySQL databases are built with a single point of failure, meaning if any component – such as drive, motherboard, or memory – fails, the entire database will fail. As a result, you might be experiencing frequent downtime, which can result in loss of revenue. You can use sharding and slaves, but these are fragile and can’t handle large amounts of traffic. A scale-out database keeps multiple copies of your data, provides built-in fault tolerance and maintains operations despite and / or disk failures.
- High developer costs – Developers working with MySQL databases must often spend a large portion of their time fixing plumbing issues or addressing database failures. Developers who work with a scale-out database are free to instead work on developing features and getting the product to market quicker. As a result, time to market decreases and e-commerce companies are able to earn revenue quicker.
- Maxed out servers – Servers maxing out on RAM for extended periods of time, or frequently throughout the day, are key indicator that MySQL can’t keep up with business growth. Adding hardware is the quick fix, but it’s also very expensive and isn’t a long-term solution. If organizations used a scale-out approach, data can be replicated across nodes, and as transactions increase in size and amount, workload is shifted to other nodes within the database.
Wrapping up
It’s clear, MySQL has its limitations, and that given time and traffic growth, any MySQL database is bound to experience performance and latency issues. And for e-commerce websites, those malfunctions will almost certainly translate into missed revenue.
After all, it shouldn’t come as much of a surprise then that a technology that was built two decades ago is struggling to keep up in today’s fast-paced digital world. Think about it: how could programmers in 1995 foresee just how powerful the Internet would actually become?
