Game-Changing Database Advances from Google Cloud, Oracle, and Databricks
And a new job that keeps me close to 'challenger' tech companies.
Hello again everyone, and a special welcome to our many new subscribers! This is my first blog post in a while, so a few things to catch up on…
I’ve been on the road the past few weeks. I traveled to San Francisco to visit my son (a software engineer), and then to the Columbia River Gorge in Oregon to hike and visit. I’m sharing a few photos because, let’s face it, who wants to see another schematic on database architecture!
Also, I have some career news: I recently started a new, full-time role as a tech writer and editor at Method Communications, an innovative PR and marketing agency, which is an exciting next step for me.
Many of Method’s clients are challenger and disrupter tech companies, such as Beans.ai, Confluent, Domo, and Freshworks. In this new role, I will continue to be very involved in the world of startups, enterprise tech, and data management, helping our clients tell their stories of innovation and differentiation.
Please follow me on LinkedIn if you want to keep up with all of that.
I’m happy to say that I will continue publishing the Cloud Database Report. Subscriptions to this newsletter grew more than 50% last year, and more people sign up everyday. This tells me that business and tech people are interested in what’s happening in the world of databases and data management—and they value the kind of independent analysis that I try to provide.
Thanks for your continued interest and support. Here’s my latest report on the database industry.
Google Cloud goes hybrid
Is AlloyDB poised to become the next big thing in database management systems? It feels that way, as Google Cloud continues to build out AlloyDB into a multi-workload and, now, a hybrid, multi-cloud database platform.
Google Cloud first introduced AlloyDB, its newest PostgreSQL-compatible cloud database, in May 2022, and it’s been adding new new capabilities over the past 11 months. Last October, the company announced Database Migration Services for AlloyDB, making it easier to migrate other Postgres databases—including those from AWS and Microsoft Azure—to AlloyDB on Google Cloud.
Now, Google Cloud has introduced AlloyDB Omni, a downloadable version of the database that runs on premises. Google Cloud says AlloyDB Omni will run “in your data center, on your laptop, at the edge, and in any cloud.” At this point, Omni is available as a tech preview.
Andi Gutmans, Google Cloud VP and GM of databases, explains that AlloyDB Omni can be used by customers who need to store and manage data on premises due to regulatory or data sovereignty requirements, or in edge deployments. Google Cloud will also offer AlloyDB Omni as a hosted database for anyone who requires that their database be isolated from the workloads of others.
It’s notable that Google Cloud, in providing a downloadable, on-prem edition of AlloyDB, is acknowledging that database services in the cloud don’t satisfy all customer requirements. And it lends credence to the argument that some organizations have slowed, or even pulled back on, their cloud deployments, due to rising/unpredictable costs and other factors.
Bottom line: Google Cloud now as a hybrid architecture for AlloyDB, and that’s a good thing. Below is my original blog post on AlloyDB.
Oracle targets developers with 23c
Oracle’s 44-year-old database keeps on chugging. Larry Ellison’s company has introduced Oracle Database 23c Free, a developer release that sets the stage for GA at some point in the future. The company announced 23c beta at its CloudWorld event last October.
There’s a long list of new functionality, including JavaScript Stored Procedures, JSON (JavaScript Object Notation) Schemas, operational property graphs, and other capabilities that are more meaningful to database developers than to business people.
Oracle is making a point of the fact that it’s giving developers first access to 23c, which makes sense. When I worked at Oracle (from 2013 to 2018), we made a major push to appeal to the new generation of cloud-native, “polyglot” developers—those who work across platforms, tools, and languages.
One new capability that is getting attention is JSON Relational Duality, which is like a toggle switch for developing in either relational or JSON/document formats. It’s yet another way for Oracle to offer customers an in-house alternative to MongoDB and other document databases.
One analyst went so far as to describe JSON Relational Duality as “perhaps one of the most important innovations in information science in 20 years.” That seems like a stretch, but it’s clear that Oracle continues to drive innovation with its flagship platform. FYI, below is a previous post on an another Oracle relational/JSON solution.
Databricks leaps from Manufacturing to generative AI
A few weeks ago, Databricks announced Databricks Lakehouse for Manufacturing, which is exactly what it sounds like—a data lakehouse for manufacturing companies.
As the name implies, this lakehouse offers industry-specific capabilities for things such as predictive maintenance, digital twins, demand forecasting, and IoT analytics. All of which could be applied in manufacturing operations.
But that’s already old news because Databricks more recently launched Dolly, a large language model that thrusts the company into the generative AI frenzy. Developments are moving fast, and Databricks is wasting no time. It introduced Dolly 2.0 just two weeks after the original Dolly.
What does Databricks bring to this fast-moving space? Dolly 2.0 is open source, and the training code, dataset, and model weights behind it are suitable for commercial use, according to Databricks. And, similar to OpenAI’s ChatGPT, Dolly 2.0 is “instruction following,” which means it interacts with user input.
Dolly 2.0 is derived from a human-generated instruction dataset. The Dolly 2.0 dataset—comprised of 15,000 prompts and responses—was generated by Databricks employees as part of a crowdsourcing exercise.
It’s worth remembering that Databricks describes itself as more than just a data lakehouse. It’s a “unified platform for data and AI.” Dolly 2.0 is another step in that strategic direction.
Further reading—and hiking
My last blog post, on booming data, was viewed thousands of times—in fact, there were more page views than Cloud Database Report subscribers, which means people shared it. In case you missed it, here’s another look.
That’s all for now. I will close with a photo from one of my hikes in Oregon’s beautiful Willamette Valley.
Public Service Announcement: Remember to get outside and take a walk as often as possible. Take care everyone!
