Escape the legacy of database costs by migrating to PostgreSQL and achieving data nirvana

Today’s CIO must plan around two truths when developing a database strategy:

• Every day, operational technology (no human interaction) and information technology (system and human interaction) create infinite data. The volume will never stop increasing.
• The cloud is the platform to create, manage, and deliver data and applications.

The need to interpret, analyze, and action data is reshaping how CTOs, CDOs, and CIOs think about their organization’s data fabric and how it supports a growing need to support cloud-native applications and data architectures. Exhibit 1 shows that the database is the top workload executives are moving to public cloud platforms.

Exhibit 1: Data requirements (database, analytics, compute, and storage) make the top workloads being moved to the public cloud providers

Sample: HFS Pulse 1H2021; n=800 Global 2000 decision makers
Source: HFS Research, 2022

Oracle is the elephant in the room

Most CTOs, CIOs, and Chief Data Officers rely on legacy SQL-based relational databases to store, query, and manage information essential to running their organization. The databases are crucial in delivering unique and timely experiences to employees, customers, and partners. In HFS’ 2022 Pulse survey of 600 decision makers, 27% cite a priority to use data as an asset as key to their corporate leadership agenda.

The leading provider of enterprise-class relational database software continues to be Oracle. At its core, Oracle DBMS delivers online transaction processing (OLTP), data warehouse (DW), and data workloads for companies to run the business, capture and format data, and transform data for AI, analytics, content management, and more. Oracle’s DBMS has been around since 1979 and has grown in complexity, robustness, and cost.

Regardless, the enterprise data lead must ask, “Does my organization need Oracle when we are developing cloud-native applications that are composable, functional, and microservice oriented?” As these applications prioritize agility in development, flexibility in architecture, and multi-cloud integrations, the need for large, legacy SQL-based databases may be nearing its end. Savvy buyers are looking to their future database to be integral to a cloud-native data fabric.

The costs of SQL, expansion, and licenses and the emerging choice of open-source, cloud-native alternatives

While there is a growing interest in NoSQL databases, most applications continue to be designed to work with relational data sources that call on SQL’s well-known extract, transform, and load (ETL), as these are proven to manage data residing across multiple locations. But the cloud is an unstoppable force of change in today’s technology and business landscape. CTOs and their peers dealing with cloud solutions and an increasing number of solutions seeking to operationalize data-as-a-service mean limiting your firm to legacy SQL solutions like Oracle’s DBMS may not meet your business users, data scientists, or technology teams’ needs.

Another reason to reconsider Oracle is the growing lock-in as Oracle continues to expand beyond data into applications, middleware, development tools, and cloud services. In addition, it is known for its licensing practices requiring customers to increase their investment and annual expenditures on Oracle solutions, even if the feature or value is unclear. This changing landscape has given rise to open-source database solutions like PostgreSQL.

Released twenty years after Oracle’s first DBMS, PostgreSQL has grown up in the cloud. With a community supporting the development of solutions around the common web-native application and data development tools like JSON and Python and multi-version concurrency control, this solution may be the key to reigning in Oracle costs without sacrificing functionality, robustness, or capabilities.

PostgreSQL, Azure, and scaling for a cloud-native operation

Unlike Oracle’s Cloud, Microsoft’s Azure cloud is a mainstay for organizations of all sizes to run mission-critical workloads, development and host applications, and secure data at rest and in transit. With more workloads moving to the cloud, it has become a natural extension of a firm’s data fabric.

As more firms adopt hybrid-cloud infrastructure, balancing data and workloads across databases is essential. But running Oracle on Microsoft to take advantage of Azure isn’t done. Further, adopting data tools like Hadoop and NoSQL requires that a hyperscaler provide tools to facilitate the flow of data across non-relational, relational, analytics, transactional, and on-premises data. Thus, the door opens for many organizations to consider an open-source solution like PostgreSQL.

For organizations running Oracle in their core data center, PostgreSQL is a destination for many data workloads as its PL/pgSQL procedural language is similar to Oracle’s PL/SQL; thus, the skills of existing Oracle DB teams can be ported to the lower-cost, open-source solution.

Automate the Oracle to PostgreSQL journey to expedite conversion and reduce costs

The similarity of PostgreSQL PL/pgSQL to Oracle PL/SQL has made migrating data and workloads easier. In addition, by using PostgreSQL on Azure, the virtualization of database resources, it is possible to lower the costs of dedicated systems, resources, and licenses of Oracle that are often dormant for a fair percentage of the time.

With data workloads more efficiently shared in a hybrid (on-premises and hyperscaler) model, more of the compute transactions needed by microservices and edge applications no longer require the spinning up of the costs of a full Oracle DB or DW. As there are often licensing and usage costs for on-premises and Oracle Cloud usage, the savings add up quickly. Further, by moving data to the cloud with Azure and PostgreSQL, an organization’s development team can focus on the development, delivery, management, and optimization efforts that bring business value and improve employee and customer experience.

While application modernization is a hot topic, data is the key. As Exhibit 2 shows, moving data to the cloud creates a path for a complete data modernization service value chain for customers and partners to prepare data for future applications and business states. Getting data right drives better, more informed decisions.

Exhibit 2: Moving data to the cloud is the start, not the end, of your organization’s data journey

Source: HFS Research, 2022

But wait, will it cost a lot to migrate to PostgreSQL from Oracle?

Moving from Oracle to PostgreSQL can be a daunting proposition. Migrating data, tables, data schema, and blocks is complex. Often this is where migration projects hit a stone wall as the manual efforts, resources, and costs appear to diminish the near-term benefits. However, software and services vendor Newt Global has developed the Database Migration Acceleration Platform (DMAP), and partnered with Microsoft’s Azure team with a solutions to show how automation Oracle take out can be achieved without overcommitting resources or conversion costs.

The DMAP platform addresses potentially finicky migration issues by automating the assessment of existing database schema to accelerate the conversion process. Any interesting aside is that by running the assessment, the software maps 50%¬–70% of the conversion and is one click away from running the post-assessment. In other words, the assessment has already done most of the work!

Nonetheless, converting a database is risky and should not be taken lightly. As Exhibit 3 shows, as part of this journey, it’s worth having a decision matrix to consider the options across data and applications and the implications of data modernization.

Exhibit 3: Consider using a decision matrix to plan your migration from solutions like Oracle to PostgreSQL

Source: HFS Research and Newt Global, 2022

Microsoft has found a partner to bring data into a PostgreSQL and Azure world

HFS interviewed Microsoft’s Azure team about its investment in automating Oracle to PostgreSQL on Azure. Their team has licensed Newt Global’s DMAP solution as part of its Oracle Take Out services to convert customers using Oracle. Since adding DMAP to the Azure marketplace, the solution has been adopted by a leading US-based retail firm, a North American personal insurance firm, and a large financial services firm. These firms have not replaced their Oracle stack, but they cite achieving significant savings through a reduction in Oracle DBMS instances and licenses.

From a technology perspective, it is worth considering the target architecture of a PostgreSQL migration on Azure. As a large telco offered in our discussion, “Our Oracle DMBS costs are multiple seven figures a year. If we can reduce 30% of our Oracle costs by migrating to PostgreSQL, are savings will more than pay for our efforts.”

Oracle isn’t the only SQL database being rationalized as organizations look to PostgreSQL. Legacy SQL from IBM, Sybase, and Microsoft can be converted to PostgreSQL based on Azure. This cloud-centric data and application model can release resource efforts from managing on-premise solutions to developing solutions that drive business outcomes.

The illustration in Exhibit 4, from vendor Newt Global illustrates an example of the time and effort savings that are possible using an automation tool to convert Oracle DMBS to PostrgreSQL using DMAP. While Newt Global provided the following as part of real-world examples, we have spoken with both Microsoft and customers about real-world experiences. And while they varied from 40%–60% of a DBMS instance being migrated, they all confirmed that equated to significant savings in resources, time, and effort.

Exhibit 4: Automating the discovery and conversion won’t move every table and schema, but it can significantly reduce the time and effort for more complex tasks

Source: Newt Global, 2022

Five steps to escaping a legacy database and embracing a more dynamic data fabric

Consider the following five questions when determining whether a database migration from Oracle DMBS to PostgreSQL on Azure is a good fit:

1. What are the technology-driven outcomes of migrating data from on-premise DBMS to cloud-based?
2. Is there a preferred hyperscaler partner for your cloud-native data fabric needs? And what is its Oracle Take out strategy?
3. What percentage of data and workloads would benefit from migrating off Oracle and onto cloud-based SQL offerings?
4. Do you have a business case (along with your technology case) for migrating workloads closer to the applications or processes used by business units or users?
5. Have you engaged your internal team and partners in developing microservices to outline current and future data and workload requirements?

The Bottom Line: Converting legacy databases can be complex and risky, and the fear of failure often leads to inaction. However, Newt Global and Microsoft have developed a strong partnership that can convert these cumbersome, legacy databases into a more dynamic, open-source solution ready for many of today’s modern applications and workloads.

Migrating 100% of your Oracle databases to PostgreSQL may not be possible. However, not exploring how to reduce your dependency on proprietary, legacy data solutions costs your firm time and money. Oracle’s database products and licensing models are evolving for the cloud, but many firms can find themselves strapped by high ‘per core’ licensing costs associated with maintaining and increasing Oracle-based data solutions as the needs for data continue to grow.

By saving 40%-60% of the time in legacy DBMS conversion, an organization’s technology team can focus its resources on addressing customization, data cleaning, or rearchitecting out-of-date data schemas to ones best suited for microservices application architectures. The automation step of migrating data to the cloud to support SaaS, custom, and workload integrations can have a positive impact on the bottom-line as they can save on financial and talent costs.