SAS Deploys and Operationalizes an Enterprise Postgres Database With Help From EDB

About SAS

SAS is the leader in analytics and the largest independent vendor in the business intelligence market. Through innovative analytics and artificial intelligence, SAS empowers and inspires customers around the world to transform data into intelligence. SAS serves customers in 150 countries, with its software installed at more than 80,000 business, government, and university sites – 88 of the Top 100 of the 2021 Fortune 500 list are SAS customers or their affiliates. 

SAS has a 40+ year history of analytics innovation and a deep commitment to curiosity, enabling employees and customers to make bold new discoveries that drive progress. This includes ensuring that teams have access to the tools and resources they need to foster creativity and problem-solving without limitations.

Challenge

SAS supports a wide range of environments under the SAS Cloud umbrella. To empower new customers and simplify their path to SAS cloud offerings, the organization launched an initiative to offer free trials of SAS products within the SAS Cloud. By making it easy for prospective customers to quickly create accounts, select products for trial, and spin up new instances within minutes, SAS anticipated increased adoption of its market-leading solutions.

One of the critical components to this strategy was the development of a new automated login and registration system for these new prospects that could scale and efficiently record contact information, desired applications and other necessary data on-demand and pass it along to the necessary systems to spin up those trials in the SAS Cloud. 

The Research and Development division at SAS  standardized on Postgres as their preferred database due to its flexibility, extensibility, open-source nature, and cost. That said, at the time of this initiative, the public-facing SAS Cloud relied principally on a mix of commercial databases. So, while it was no surprise that the new login system relied on Postgres as its backend database, the SAS Cloud team had limited Postgres-specific database expertise on staff to help support it. 

Meanwhile, the accelerated trial program had an expected delivery date of a mere two weeks to stand up a mission-critical, highly-available, customer-facing Postgres database to support an enterprise-class production-ready instance of this newly developed application.

David Wagoner, Data Platform Engineering Manager at SAS, remembers, “We had just two weeks to get this system ready, and my team had never deployed Postgres before—it just wasn’t in our team’s service catalog at the time. I was the only one on my team that had used Postgres before, and I had enjoyed working with the database, but I needed to get my team of Oracle and SQL Server DBAs ready to operationalize, manage, and support a production Postgres system in a matter of weeks.”

Solution

Driving success with enterprise support and on-site training

The SAS Cloud team provides 24/7 support for their platforms and as a result is accustomed to working closely with database vendors, taking advantage of enterprise-level support packages to ensure access to hotfixes, bug fixes, troubleshooting support, and immediate response when needed. They knew that they would need a similar relationship with a Postgres support vendor, as well as someone that could provide architectural guidance and training as the team came up to speed on Postgres. 

Wagoner explains, “We knew we needed a Support Partner that could bolster our Postgres expertise, train our team on how to make the most of Postgres, guide us toward the best tools to support the database, and provide on-call support for troubleshooting. After looking at the options, it was clear that EDB could meet our needs.”

Despite the short training window available, EDB was able to identify a talented Postgres expert to come on-site and train not just the database architects on the team, but also members of both the Tech Support team and the Research and Development organizations.

“We were able to get a large number of people familiarized with the specifics of Postgres, which enabled our team to start looking at Postgres for other use cases as well,” continues Wagoner. “By getting a strong foundation through training, we could build the in-house skills to use Postgres for other projects. EDB's online, self-service training has helped as we’ve expanded our team, ensuring that we don’t need to wait for a critical mass before engaging in advanced training.”

On the support side, one distinct advantage that EDB has is the access to a deep bench of Postgres experts intimately involved in the project. Wagoner recounts one memorable support case where they had a tricky database recovery. The support case was escalated, and the team ended up getting access to a frequent committer on the EDB team who was able to help resolve the issue.

Operationalization enabled by trusted tools

“We had an aggressive timeline, and we knew we needed help not just with training and support, but also with identifying the right tools to support day-to-day operations,” describes Wagoner. 

In particular, Wagoner points to tool selection as an area where he and his team saved a significant amount of time by relying on EDB. Given the on-demand nature of the application, Wagoner and his team needed to architect an enterprise-class environment, complete with enterprise-class tools for replication, monitoring, high availability, failover management, backup, recovery, connection pooling—in short, all the pieces required to operationalize a global, always-on system. 

If the team had evaluated the universe of tools available in the Postgres community, they would have needed to spend significantly more time and effort on due diligence testing and operationalization of those tools. 

“I know the EDB team has already vetted a subset of tools and can provide me with the relevant reference architectures, installers, patching automation, etc. that I need to feel confident,” says Wagoner. “It really eases the burden of ensuring the environment is architected well.”

With the pressure on for a fast turnaround, the simplicity of using EDB to cut through all the options and isolate a vetted set of tools that would meet the needs of the SAS Cloud team was invaluable. Wagoner highlights Postgres Enterprise Manager as a linchpin to the database’s day-to-day operations. 

Result

The combination of training, services, support, and tools made what seemed to be impossible a reality. Within a matter of weeks, the team had a brand new, custom-built, mission-critical application up, running, and highly available, complete with a team trained to support it. Wagoner and his team took full advantage of the embedded capabilities and installers available with EDB Postgres Advanced Server, which helped the Oracle-familiar team transition more easily, as well as EDB Postgres Enterprise Manager for monitoring, management, and optimization. By taking advantage of additional tools such as EDB Failover Manager and EDB Backup and Recovery Manager (BARMAN), the team also ensured that the system would be appropriately protected against unexpected issues.

Since that first customer-facing, production database, SAS has significantly increased its use of Postgres throughout the organization (beyond R&D), with databases that range in size from tens of gigabytes to multiple terabytes of data. 

In addition, enterprise-class Postgres is now a data platform option within the SAS Cloud for customers looking to do analytics on it. With the security embedded in SAS Cloud, customers can be certain that their sensitive data is protected and hosted in a secure system within SAS Cloud.

Wagoner points out, “The Postgres community is constantly using it in innovative ways. You go to Postgres conferences, and you get to meet so many people doing cutting edge work. Since its open source, the Postgres community really takes the lead on new technologies and approaches, improving speed to market and time-to-delivery. I think that’s why Postgres was one of the first to be a viable, containerized database—to really be Kubernetes ready. We were doing deployment automation in 2015, and I don’t think we could have done that as easily with another database at that time.”

As SAS’ use cases have expanded, the team has gone back to EDB for expert help through their professional services offerings. In one instance, they had an application that needed to support OpenStreetMap, so they could overlay geographical maps on some of the company’s statistical software. Since OpenStreetMap uses the PostGIS extension, the team tapped EDB to provide an expert that specialized in PostGIS implementations.

Wagoner cites other examples of where Postgres shines as well: “We have a development team working on network automation. Postgres natively supports a few networking data types that other databases don’t support out-of-the-box. When it comes to extensibility, Postgres really stands out. Just consider the number of languages supported, the number of index types, the JSON capabilities. Postgres has been a leader in new feature development.”

Future

The SAS Cloud team continues to identify new use cases for Postgres and take advantage of the latest advancements championed by EDB. In addition to expansions of the existing implementations, Wagoner also rattles off active investigations of EDB Postgres Distributed, EDB Postgres for Kubernetes, and EDB’s fully managed offering, BigAnimal, for near-term projects. 

“EDB has provided us the Postgres expertise, tools, documentation, training, professional services, and technical support required to meet the enterprise-class requirements of SAS Cloud,” concludes Wagoner. “While it’s certainly possible to self-manage Postgres without EDB, if you’re trying to manage a large-scale database or a mission-critical application, you’ll end up spending more money in time, resourcing, and staffing to do it yourself. With EDB, you get a ready-made product that’s packaged and supported by an enterprise vendor.”

Want to learn more? Here’s how to get started with EDB. 

Do you have an EDB Postgres success story you want to share? Get in touch with us via our website or at success-stories@enterprisedb.com