By Patrick Moorhead and Azure Security News
As noted in its recent earnings call, Oracle Autonomous Database has been growing rapidly, increasing 55 percent year-over-year in constant currency in the most recent quarter, Q3FY21. And Autonomous Database continues to innovate without a direct response from its competition—after nearly three years in the market.
I am not aware of any other fully autonomous, machine-learning powered, self-driving, self-securing, self-patching, cloud data warehouses available from any other vendor—be it AWS, Snowflake, Microsoft Azure or the more legacy database vendors such as IBM, SAP and Teradata. The competition has been very busy creating focused, single-purpose databases. AWS, for example, currently has 16 of them, and counting. Strategically, I get this. I once had a “real job” running strategy, have run billion dollar businesses, and if I had to compete with Oracle, this is what I would do. You always attack “integrated” with “best of breed”. I believe the enterprise challenge becomes integrating the disparate databases to work together.
As of today, there are several variants of the Oracle Autonomous Database portfolio, including Autonomous Data Warehouse (ADW), Autonomous Transaction Processing (ATP), and Autonomous JSON Database (AJD). Unlike other vendors’ single-purpose, focused databases in the cloud, Oracle Autonomous Data Warehouse supports multi-mode and multi-workload requirements, within a single converged database engine, including operational, analytic, JSON document, graph, ML, and blockchain. There is no ETL required to move data, no separate services to invoke, as everything is engineered into the same database. This is different from going from AWS Aurora to AWS Redshift, where customers invoke a Lambda function, followed by a Kinesis function, to a S3 bucket and then load the data into Redshift. In Oracle Autonomous Database, I believe transactional data can be combined with analytical data with relative ease.
There are thousands of Autonomous Database customers globally, including companies relying on it for data warehousing such as ride hailing innovator Lyft, $50B global insurance leader AON, telecommunications leader Vodafone, content provider and broadcaster Sky, the largest insurance company in Italy and one of the largest worldwide, Generali, as well as Accenture, Outfront Media, MineSense, real estate conglomerate Kingold, financial services provider Certegy, and many more.
My conclusion? Many customers are voting with dollars for an autonomous cloud data warehouse solution rather than with Snowflake, grappling with manual labor-intensive administration and management. Does anybody actually wake up and eagerly anticipate spending more time and money on the cloud than is really necessary? Does managing 10 API calls instead of one make anyone feel better about themselves? Likely not. This is what Snowflake customers grapple with.
Today, Oracle announced the next generation of its Autonomous Data Warehouse, designed to make it so easy to use that even entry-level enterprises, SMBs and companies without much IT expertise can now load data, transform data, cleanse data, as well as automatically create business models—and automatically discover patterns to generate insights. I believe Oracle is expanding its addressable market by making its Autonomous Data Warehouse so intuitive that it’s becoming more broadly applicable to a much wider range of organizations, vertical industries, and users. Think that new fashion startup without a dedicated IT department can’t figure out how to fire up a cloud data warehouse to support its booming business model? Think again.
With today’s news, I see the same philosophy applied as in the autonomous database—use automation to eliminate a large proportion of the administrative tasks resulting in a point-and-click and drag and drop user experience.
Let’s dig into the new features and give examples of how they will improve various people’s lives within your organization.
No IT Support Required
Cloud data warehouses today are designed for technical users such as data scientists and data engineers. Oracle is taking a different approach and using automation to deliver a point-and-click, drag and drop experience that’s so intuitive it’s like the iOS of the enterprise cloud data warehouse space.
Just as cloud HCM systems today are self-service, Oracle believes the data warehouse should be no different. Oracle makes life easy for business professionals by providing a self-service data warehouse. To work with data and not require them to know SQL, which is how every other cloud data warehouse works today. This is analogous to what Oracle has done in the HCM space—employees of any company using Oracle HCM can simply have a conversation with their phone to submit and check on expenses, order new business cards, check on shipments and open requisitions. This intuitive approach has replaced manual labor and physical departments solely once dedicated to these functions.
If you are a data analyst, then you are in luck as Oracle made you the primary focus of the enhancements. They are the prominent people who work with the data in the data warehouse. A lot of time, they spend cycles trying to convince IT support to move data from one location to another.
Oracle has given thought to the complete life cycle of data. The ability to load data, do data transformations and build a business model.
To zero in on data loading. Today to load data in your data warehouse, you need to create a table. To do so, you need to know SQL and understand relational tables. Loading the data requires the proper credentials and the correct data formats for your files.
Oracle has realized this IT-centric solution is not what most businesses want or at least the actual users of the data.
In Oracle’s next generation cloud data warehouse, a user interface (UI) and included tools easily load data. While I haven’t personally logged in, the videos look as if it were as simple to use as logging into XBOX Live. With the proper credentials, of course. It provides the ability to load data from anywhere, even other clouds, using a simple drag and drop interface. The system previews the data, decides what the data type should be, lets the user create a new table (without any SQL commands), and provides a preview of what it’s going to look like when it goes into the database.
I believe everyone who works with data will appreciate what Oracle has done here with a UI and an embedded data transformation capability along with filters and data cleansing operations.
Oracle also has embedded analytic capabilities deep into its database. Machine learning (ML) algorithms will look at the data loaded and provide suggestions for your business model. There are a broad array of data sources available such as Oracle Fusion Applications, SAP, Salesforce and more.
Oracle provides the ability to do business modeling. With multiple data sets, there are various tables. Oracle will automatically identify the relationship between the data sets providing a business-level view of how the business analytics tool will query the data.
As an example, think of an online movie streaming company. Oracle’s next gen cloud data warehouse might identify from your data load that the sci-fi genre for the month of June had much fewer viewers or fewer purchases than forecast or expected, a fact previously unknown.
Oracle does recognize that users have preferred third-party tools such as exporting data into Excel or visualization using Tableau. For the data analysts, Oracle has tools and a choice to use other third-party tools. Oracle Autonomous Data Warehouse is open, and can be used with data and tools in other clouds, including AWS and Azure.
Citizen Data Scientists
Oracle is also looking to help data scientists, especially citizen data scientists, those whose primary role is outside the field of analytics.
Following the same principles, Oracle provides the same load and transform capabilities. Additionally, an intuitive interface makes it much easier to build a machine learning model.
Data scientists have questions such as “What is the likelihood of this customer purchasing a product?” or “What is the likelihood of this customer leaving my service?” These are all questions that machine learning can now answer.
Typically it is done by loading historical customer data and building a model that predicts what customers will do next.
Oracle has done a lot to automate this process with a technology called AutoML. Using a machine learning landing pad as the starting point, data scientists use AutoML to build a model. It is then a case of picking a data source and defining what you are trying to predict.
For example, it could be a retailer trying to predict the customers who are likely to say “yes” to an offer for an affinity card.
Under the covers, AutoML is a well-defined data science innovation with over 20 built-in algorithms within the database. The AutoML algorithm chooses the right candidate algorithms based on the data, selects the right attributes to feed into the algorithms, and corrects parameter settings to tune those algorithms. It essentially sets up a competition between the different machine learning algorithms to identify the best algorithms for your problem.
Once the data sets are prepared and the machine learning models are built, guided by AutoML, deployment is via REST. In contrast, there are no built-in ML tools in AWS Redshift or Snowflake—these are all external services, some provided by third parties, with separate billable rates and contracts, adding complexity and security challenges with the constant need to move data across isolated databases.
I saw the demo and to be honest, I didn’t believe what I saw, asked around 50 questions, but do believe it’s real.
Line-of-business (LOB) developers are the third category of users looked at by Oracle in its next gen Autonomous Data Warehouse announcement. The distinction here is not the back-end Java developer but the developer who needs a quick application. They need to build something fast, using a low-code environment.
LOB Developers get all the new features we have discussed and some existing tools available for about a year now such as APEX.
Oracle APEX (Application Express) Application Development Service is an integrated, low-code development environment entirely built into the Autonomous Data Warehouse. It’s an end-to-end application development environment, using low-code techniques to build, deploy and manage applications. There are over 500,000 APEX users worldwide today, and many of the best looking modern apps have been built on APEX, including Oracle’s Public Health Management Applications Suite, currently in use by the Centers for Disease Control (CDC) in the U.S.
APEX at MineSense
APEX is also very popular with small companies. One such example is MineSense, who provides a service to physical mining companies. MineSense does not manage and run the mines, but instead provides a service to mining companies by instrumenting the excavators. X-ray sensors are installed on the shovel, measuring the purity of the ore excavated. The data warehouse provides a map of the mine and identifies the location of the most valuable ore.
An application using Oracle APEX runs on a tablet in the excavator’s cab. It specifies in real-time, based upon the analysis and X-ray data, whether the shovel’s dirt is going to the processing pile or the refuse pile. This example represents the perfect blend of new tech forming synergies with old tech to propel both businesses into the future.
Many vendors now offer a cloud data warehouse, but few are trying to accomplish what I believe Oracle’s next gen Autonomous Data Warehouse is trying to do. With Oracle’s cloud data warehouse, customers do get a single, converged cloud database that should meet their business requirements, removing the need to buy specialized databases and integrate data from multiple sources. I could be swayed by “best product” evidence, but this still doesn’t comprehend the enterprise’s need to get a total view of all the data or the work involved meshing disparate databases together.
Table stakes in the cloud data warehouse market is a SQL-centric data warehouse. What is exciting about this announcement is that Oracle is taking the market in a new direction for cloud data warehouses, making them useable by virtually anyone in any size organization by introducing degrees of simplicity that prior to this announcement were completely unavailable in this space. This is the big takeaway.
Oracle set out to make the cloud data warehouse autonomous. Doing so has removed a large part of the operational administration costs and manual labor, I believe, saving money immediately, and longer-term, delivering better reliability and service levels, and now includes a better user experience with an automated system.
I do believe many kinds of enterprises see the benefits of autonomous, eliminating administration, repurposing headcount in more strategic areas. Simplification means deploying applications more rapidly with less time wrestling with infrastructure. Now I can build my data warehouse a lot more rapidly. If I want an autonomous car, I would invest in a Tesla. I wouldn’t visit Pep Boys and attempt to cobble one together. There are cloud-native companies that do and I get that. If I want a cloud data warehouse, I don’t see why I would try to cobble dozens of different databases, tools, and services together like Snowflake and hope to achieve the same result as Oracle has made available in Autonomous Data Warehouse.