The News: Oracle Autonomous Data Warehouse (ADW) is a cloud data warehouse service that is designed to mitigate the complexities of operating a data warehouse, securing data, and developing data-driven applications. It is developed to automate provisioning, configuring, tuning, scaling, and backing up of the data warehouse and includes tools for self-service data loading, data transformations, business models, automatic insights, and built-in converged database capabilities that enable streamlined queries across multiple data types and ML analysis. ADW is available in both Oracle Cloud Infrastructure (OCI) and customer’s data centers with Oracle Exadata Cloud@Customer, and Oracle Dedicated Region Cloud@Customer, enabling customers to meet data sovereignty requirements. Read the Oracle ADW link here.
Oracle Autonomous Data Warehouse Rains on Snowflake’s Cloud Data Platform Parade
Analyst Take: To fully understand the competitive benefits of the Oracle Autonomous Data Warehouse (ADW), an overview of the Oracle Autonomous Database portfolio is warranted. The technological vision that drives Oracle Autonomous Database (ADB) portfolio development emphasizes minimizing and the elimination of time-consuming, error-prone manual administration of DB systems and operations. Through fulfilling this portfolio development vision, Oracle ADB enables organizations to automate legacy manual processes and improve reliability and security by mitigating human error.
Oracle ADB brings together full infrastructure integration, comprehensive DB automation, and automated data center (DC) operations to provide a complete cloud DB service to customers. Key to the success of ADB is delivering use case optimization of the applications most prioritized by customers consisting of Autonomous Data Warehouse (i.e., analytics, data science, and machine learning), Autonomous Transaction Processing (i.e., business applications, analytics, and mixed workloads), and Autonomous JSON Database (i.e., JSON Document Management, with a pushbutton upgrade to ATP).
As a result, Oracle ADB, delivered three key breakthroughs in the areas of (1) fully autonomous, no administration DBs; (2) cloud elastic model that assures customers pay only for what they use; and (3) broad platform capabilities that delivers a single converged DW service.
Oracle ADB Delivers the Joy of Fully Autonomous Database Capabilities
Architected from the start on cloud design principles, I see Oracle ADB as providing the highest priority capabilities needed to comprehensively meet fast-evolving customer demands. These include:
- Auto-provisioning: Automatically deploys mission-critical DBs which are fully secure, fault-tolerant, and highly available.
- Auto-configuration: Automatically configure the DB to optimize for specific workloads.
- Auto-indexing: Automatically monitors workloads and proposes index change that could accelerate applications.
- Auto-scaling: Automatically scales compute resources online when needed by workload.
- Automated Data Protection: Automatically safeguard sensitive and regulated data in the DB, all through a unified management console.
- Automated Security: Automatic always-on encryption for the entire DB, backups, and network connections.
- Auto-Backups: Automatic daily backup of DB or on-demand.
- Auto-Patching: Automatically patches or upgrades with zero downtime.
- Automated Detection and Resolution: Using pattern recognition, hardware failures are automatically predicted without long timeouts.
- Automatic Failover: Automatic failover with zero-data loss to a standby database.
With fully autonomous Oracle DB functions, organizations can rely on highly available, self-healing infrastructure due to triple-mirroring of disks designed to minimize any fallout from disk failures, simple disaster recovery, and online patching for complete protection. Additionally, organizations lock-in secure DB infrastructure by adopting always encrypted, always audited, and always patched features, as well as automated data protection that includes risk analysis and assessment of user privileges.
Cloud Elasticity Means Pay Only for What You Use
Oracle ADB enables organizations to right size their data workloads according to the exact number of OCPUs and TBs of storage they require and grow granularly as workloads increase. As such, they are not constrained by fixed building blocks or “t-shirt” sizes offered by rivals like Snowflake. Moreover, Oracle ADB auto-scales instantaneously up or down with no downtime or manual involvement by adjusting CPU (Central Processing Unit) and IO (Input Output) resources based on immediate workload requirements. In addition, with Oracle ADB organizations can also realize savings by shutting off compute functions for idle systems and, equally important, also have the confidence to restart them instantly as needed.
Oracle ADB’s Broad Platform Capabilities that Deliver One Converged DW Service
As a result, Oracle ADB augments the business value of analytics, especially built-in and fully integrated SQL analytics, throughout Autonomous Data Warehouse (ADW) implementations. This includes the full leveraging of ML for discovery of hidden patterns and actionable new insights, spatial analysis to discover interactions based on geographic relationships, and new graph analysis used to discover related connections and patterns.
Providing the Direct Benefits and Deployment Flexibility of the Converged Data Warehouse
Oracle ADW natively includes dozens of ML algorithms, particularly including in-DB algorithms and analytics functions with SQL and Python APIs. Oracle ADW performs parallel ML directly in Data Warehouse or OLTP for fast model building and real-time scoring of new data. In addition, the platform explores and prepares data, builds and evaluates models, scores data, and deploys solutions as well as keeps data secure by avoiding copy contagion and corruption as copies of data in analytics systems are a common source of data breaches. ADW also include Application Express (APEX), a highly productive, no-code/low-code environment favored by developers globally.
Moreover, Oracle ADW now supports self-service tools for Data Analysts, including simple drag and drop loading, declarative transformations and data cleansing, automatic creation of powerful business models, and guided discovery of hidden patterns and anomalies. Further appealing to the demands of DB customers, ADW is open, supporting robust integration with third-party tools and built-in loading from other clouds such as AWS, Azure, and Google. Specifically, cross-cloud interconnects are established between Oracle and Microsoft Azure regions providing streamlined integration of Azure-based systems with ADW.
Snowflake’s Cloud Data Platform Background and Approach
Snowflake’s Data Cloud Platform is engineered to support what it refers to as the Data Cloud and seeks to enable organizations to “unlock the value of their data.” The Snowflake architecture logically separates but integrates storage, computing, and services. The architecture aims to allow users and data workloads to access a single copy of data without impacting performance. Snowflake looks to abstract the complexity of underlying cloud infrastructure, allowing organizations to run their data across multiple clouds and regions for a single experience.
In addition, Snowflake targets secure collaboration across the ecosystem by supporting organizational access to additional shared data sets and data services via Snowflake Data Marketplace and providing similar connections with the Snowflake customers that comprise the broader Data Cloud. In September 2020, Snowflake became a public company through its IPO that raised $3.4 billion. Snowflake gained attention and hype around its IPO, including financial backing from players like Salesforce and Berkshire Hathaway.
In recent developments, Snowflake has made top-line growth strides, but its net loss expanded from $348.5 million in fiscal 2020 to $539.1 million in fiscal 2021. That trend continued in Q1 2022, as its net loss more than doubled from $93.6 million to $203.2 million. I see this as problematic. Snowflake is unenviably dependent on the kindness of strangers, since it must rent cloud computing resources from hyperscalers like AWS and Microsoft Azure, who also peddle their own cloud data warehousing solutions, which I also see contributing toward keeping price points down for Snowflake’s own offering into the foreseeable future. In contrast, I see Oracle ADW excelling in the area of cost governance, and through Oracle Cloud Infrastructure, avoids the dependencies that can limit Snowflake’s pricing and data workload management flexibility.
Key Takeaways on Oracle ADW’s Key Differentiators versus Snowflake
Taken together, I view Oracle ADW as clearly differentiated in the five key areas that organizations assign top priority – cost governance, real-time DW workloads, data integrity, ML integration, and deployment flexibility – in comparison to the Snowflake Cloud Data Platform.
First, in terms of cost governance Oracle ADW provides granular compute sizing, whereas Snowflake requires customers to double in size and cost for every step up to meet their expanding compute size requirements. In parallel, Oracle ADW auto-scales instantaneously online and uses governed storage, however I see Snowflake lacking the oversight mechanisms needed to control storage costs.
Second, I see Oracle ADW as providing a definitive edge in the processing of real-time and operational DW workloads. For example, Oracle ADW supports large numbers of concurrent queries, whereas the Snowflake platform defaults to eight concurrent queries per clusters. Oracle ADW furnishes indexes for rapid lookups – a capability that Snowflake completely lacks. In addition, Oracle ADW assures efficient updates, while Snowflake offers only an append-only architecture that impedes real-time updates.
Third, Oracle ADW ensures data integrity by applying enforced unique/primary key, foreign key and check constraints to ensure that the data is correct by preventing simple mistakes like duplicate records. In contrast, the Snowflake Data Cloud Platform does not enforce such meaningful constraints. As a result, Snowflake customers do not have full assurances about the integrity and correctness of their data.
Fourth, Oracle ADW supports and integrates a wide array of built-in, self-service ML algorithms. With Snowflake, customers must license third-party ML tools, install, manage, and learn how to use the tools, delaying time to insight and increasing overall costs. Oracle ADW also includes APEX, a popular no-code/low-code environment that significantly accelerates application development. I do not see Snowflake having equivalent of Oracle APEX.
Fifth, Oracle ADW is available in OCI and on-premises in Oracle Exadata Cloud@Customer and Dedicated Region Cloud@Customer with complete architectural identicality across deployment models. Moving Oracle Database data does not require transformations or re-formatting, Further, Oracle’s Cloud@Customer options enable customers to meet data sovereignty and regulatory requirements. Snowflake only runs in the public cloud and from my perspective does not address the requirements for data sovereignty and regulatory compliance.
Overall, I believe that DB/DW decision makers must prioritize these selection criteria in the evaluation process and and direct comparison of the Oracle ADW and Snowflake propositions. Organizations need to consider the full spectrum of DB and DW requirements or risk selecting a solution that curtails their ability to perform analytics in real-time, lacks fine-grained elastic scaling, does not provide full data integrity, has insufficient tools, and limits deployment flexibility in advancing their cloud DW journey. By separating hype from the underlying realities of DW optimization and administration, organizations can avoid spinning their wheels on subpar outcomes and getting ensnared in data warehousing snowdrifts.
Futurum Research provides industry research and analysis. These columns are for educational purposes only and should not be considered in any way investment advice.