Table of Contents
- Usage Scenarios
- Decision Criteria
- Company Analysis
- Key Takeaways
- About William McKnight
- About Jake Dolezal
- About GigaOm
Enterprises from every industry and at every scale are working to leverage data to achieve their strategic objectives—whether those are to become more profitable, effective, risk tolerant, prepared, sustainable, and/or adaptable in an ever-changing world. An enterprise’s data maturity must grow at pace with the business and its needs to achieve agility and resilience—otherwise it will be hamstrung or tripped up by limited data capabilities. A mature analytic data management strategy includes the ability to adapt.
To leverage data, it must be accessed, combined, and governed efficiently and managed effectively. As consultants and analysts, we frequently encounter organizations that struggle with their analytic platform choices.
While the pain caused by a lack of mature analytics may be related to people or processes, we also find enterprises have not adopted or utilized the right products or are stuck with outdated tools—often because the technical debt accumulated from years of workarounds and gap-fixing existing processes is too “expensive” to simply “rip and replace” with more capable modern databases. However, the demand for mature, modern data analytics is becoming too strong to ignore. Getting by with an outdated platform is no longer a sustainable choice. So, it’s time to consider analytic platforms that have evolved to meet the data challenges of modern businesses.
We considered the potential for longer-term relevancy as needs shift from data warehouses, lakes, and ETL exclusively to a combination of data warehouses, data lakes, data fabrics, AI, and pipelines.
We’re focusing on platforms that support a broad range of modern use cases, often cross-functional ones, that benefit from a comprehensive platform. Examples of these use cases are: customer loyalty and churn, predictive maintenance, fraud detection, and supply chain optimization.
For instance, many of the success factors for supply chain optimization relate to how well the supply chain process is working and the ability to rapidly predict and remediate disruptions in the supply chain. This use case requires cross-functional data (such as order management, sales, products, logistics, suppliers, and so on) from many source systems that must be accessed, integrated, mastered, and governed—sometimes in real time—adding to the scope and complexity of achieving a modern, mature, networked supply chain.
We are considering a vendor’s complete analytic offerings, which in some cases are multiple tools, as the platform. We are evaluating these platforms based on decision criteria described in this report.
The correct analytic platform selection is so important to these use cases that any enterprise level agreement with a vendor is just one of many factors in the selection process. In other words, even when there is an enterprise contract with a vendor, companies should consider architectural alternatives. The lack of platform fit and features leads to increased configuration and management challenges.
At the very least, you can see where you may have some gaps between expectations and outcomes in your analytics solution, which can help you budget and scope your project better.
This report is the fourth in a series of enterprise roadmaps addressing cloud analytic databases. The first two reports focused on comparing vendors on key decision criteria targeted primarily at cloud integration. Those vectors represented how well the products provided the cloud features corporate customers have come to expect. In 2017, we chose products with cloud analytic databases that deploy exclusively in the cloud, or had undergone major renovation for cloud deployments. In 2019, we updated that report with the same vendors. In 2020, we had new vendors and a new name. This year, we reviewed and adjusted our inclusion criteria.
We are now targeting technologies that tackle the objectives of an analytics program, as opposed to the means by which they achieve those objectives. These days, many believe the best vessel to be a data lake/cloud storage (not necessarily in relational format). And many are finding ways to join the relational database with the lake as a “lakehouse” or “unified analytics,” treating the data lake as external tables.
We believe no reference analytic architecture is complete without database technologies, despite the inclusion of the data lake and the lakehouse concept, in the architecture. The characteristics of so-called “big data,” for example—larger amounts of data with fluent ingestion, but with a smaller, science-based user population—make the data lake appropriate as well.
In this paper, we focus on the higher-volume, critical-app compute and storage that is the analytic database. We have undertaken the ambitious goal of evaluating the current vendor landscape and assessing the analytic platforms that have made, or are in the process of making, the leap to a new generation of capabilities in order to support the AI-based enterprise.
For this Roadmap Report, we chose technologies powered for an enterprise-class application in a midsize to large enterprise. We considered popularity and interest. The vendors/products chosen were:
- Actian Avalanche
- Amazon Redshift
- Cloudera Data Platform
- Google BigQuery
- IBM Db2 Warehouse on Cloud and Cloud Pak for Data
- Microsoft Azure Synapse Analytics
- Oracle Autonomous Data Warehouse
- Teradata Vantage