Infrastructure upgrades drive efficiency, reliability

Cue the virtual confetti.

In September, we officially powered on the final phase of the Matrix Platform infrastructure upgrade, and all environments are now running in the new infrastructure at full throttle. Since mid-May, our database team, led by Kevin Ruff, has migrated all the Matrix Platform environments — Dev, QA, UAT, SIT, Prod Support and the Production databases.

This upgrade to SQL Server’s current best-in-class data management system positively impacts the overall performance of the Matrix TMS/SCM and FSS user interface and greatly decreases the overall processing time. The new infrastructure provides a tremendous boost in processing power; 3x improvement for FSS and 2x for TMS/SCM, while also doubling the platform's memory capacity.

It also enabled us to migrate all lower environments — Dev, QA, UAT and SIT — from virtual servers to the same high-performance servers used by Matrix TMS/SCM/FSS, eliminating the gap between the test and live environments.

As an example, in May we moved UAT, the platform’s testing environment, from a virtualized server to a physical server with more than eight times the processing power. In real world terms, that's like taking it from a single-lane highway with no passing lanes to a multilane highway with fewer delays and smoother flow. This upgrade finally allowed us to create a parallel environment that mirrors the Matrix TMS/SCM/FSS production environment, making performance testing and support more reliable.

This kind of behind-the-scenes improvements can be seen in search screen performance. In each environment, users will notice better reliability, enhanced efficiency and quicker throughput.

More power, less lag

There’s no need for hype, the data already speaks for itself.

On an average day, Matrix TMS/SCM users run more than 800 million database queries.

That's a lot.

Many queries return results or modify the data in milliseconds. So, it may not be obvious that the system is performing better in a lot of cases. Shaving 12% to 75% off a query is fantastic. But most users won't truly notice the difference if their response improves from 390 to 340 milliseconds.

However, where they will notice a difference in more complex queries. On the old servers, we had about 400,000 queries a week that took over 1 second. On the new servers, that has been reduced by 26 percent.

Rock-solid reliability

The upgrade to SQL Server allows the Matrix Platform to take advantage of SQL Server Availability Groups. In non-tech speak, think of the Matrix Platform as a city that relies on one main power station (aka the primary SQL Server). If Matrix were to experience a blip, one of the backup power stations (aka the different SQL Server Availability Groups) automatically takes over. These Availability Groups are constantly updated which means the lights stay on, traffic flows and business continues as usual.

For our users, that means:

• Databases stay online: If a server fails, it automatically switches to backup to avoid downtime.
• Disaster protection: It can store up to eight copies of a database, including ones at different locations.
• Flexible failover: During outages or maintenance, it switches between servers automatically.
• Read-only copies: It uses backup servers for reporting and other read tasks to reduce pressure on the main server.
• Built-in monitoring: Windows clustering is used to check server health and manage failovers.

The road already traveled

This journey began with a technology audit completed by Wavestone in 2023, which gave us a clear roadmap: reduce technical debt, improve platform security, modernize our servers, and boost overall performance.

After two years of focused effort, the Matrix Platform’s engineering and database teams have knocked out the audit’s to-do list.

In addition to the SQL Server upgrade, three notable enhancements include:

Code Analysis

With more than 16 million lines of code, the Matrix Platform requires regular software maintenance that's critical to the platform's stability. We’re using multiple third-party tools to scan the code for potential quality and security issues to identify in-line elements that require cleanup to avoid potential bugs or future coding issues. As a result, our team resolved 3,927 outstanding items. There’s also a focus on removing any unused or dead code to continue to boost platform performance.

Component Upgrades

Stability is key to the Matrix Platform and this extends to third-party components and tools that are utilized. These components and tools routinely are upgraded to the latest versions to help resolve bugs, enhance performance issues and address security concerns. Tools upgraded included Telerik, Aspose, SiteMinder, IdentityMinder, MMER, LDAP, Tibco and more.

Web API Migration

The existing .Net remoting service is being upgraded to utilize the Web API technology, allowing the platform to send and receive larger amounts of data in less time. In non-tech terms, our move to Web API technology is similar to how each cell phone provider originally used their own individual, proprietary communication networks. Now, all cell phone providers use the same 5G service which creates better network availability and service.

“The infrastructure upgrade represents a true team effort and a major leap forward for our platform,” said Scott Biernacki, head of the Matrix and WorldPort Factory. “The seamless transition and enhanced performance lay the foundation for continued reliability and future innovation that help support CEVA's Ground & Rail and Lead Logistics businesses.”