Google Cloud Platform Technology Nuggets — December 1–31, 2024 Edition
Welcome to the December 1–31, 2024 edition of Google Cloud Platform Technology Nuggets.
Thank you so much for subscribing/reading this newsletter throughout the year. I wish you all a Happy New Year 2025.
This is going to be the last edition for the year and I decided to write a consolidated one for December, since there are not likely to be too many releases / announcements as we end the year. Google Cloud has published a few blog posts that looks at the year 2024 and various announcements vis-a-vis specific products, so you might find a few articles like that across this post.
The nuggets are also available in the form of a Podcast. Subscribe to it today.
Top Google Cloud Blogs of 2024
Its a great time to look back on the year that was 2024 and the top Google Cloud Blogs published for each month. I am not sure if all of them got covered by this newsletter but check them out.
And while on Google Cloud in general, do note that registration is open for Google Cloud Next ’25 that will be held from 9–11 April in Las Vegas. Check out the blog post. Early bird registration is now available for $999 for a limited time.
Containers and Kubernetes
The Kubernetes podcast is one of the most listened to by Kubernetes practitioners. This year, it covered several podcasts that celebrated the 10th anniversary of Kubernetes along with specific release announcements. Check out the blog post that covers the The 10 top Kubernetes Podcast episodes of 2024. Bonus in this blog post, is 2024 recap episode.
Identity and Security
An organization policy provides a centralized, programmatic control over your organization’s resources. A policy identifies a set of constraints or restriction that can apply your cloud resources across organization, folder, or project level. The custom organization policy tool has expanded in Google Cloud to cover additional 30 Google Cloud services. Check out the blog post for more details and getting started with it.
Mandiant Academy’s newest on-demand training, the cyber threat intelligence (CTI) design playbook has been released. The track contains three two-hour courses addressing different aspects of CTI program development: Designing and Building a CTI Program and Operating an Effective CTI Program. Check out the details.
The first CISO Perspectives for December is out. This edition focuses on Cybersecurity threats in the year 2025. The report mentions that we should expect more use of AI when it comes to cybersecurity threats, threats moving ahead to find more attack vectors and more. Check out the report.
Machine Learning
This month saw a series of announcements around Gemini 2.0 and related AI services. Top of the announcement was the availability of Veo and Imagen3. Veo, now available on Vertex AI in private preview, empowers companies to effortlessly generate high-quality videos from simple text or image prompts. Imagen 3, now available to all, generates realistic and highest quality images from simple text prompts. Results from both these services is fascinating. Do check out the blog post for various examples.
While the attention has shifted completely to Gemini 2.0 models, here is an excellent blog post that highlights the capabilities of Flash 1.5 Pro models to process video inputs and produce specific information that you are looking forward. The post highlights a video of a dish being prepared and how you can segment it into the various stages like preparation, plating, etc and identify the positive and negative moments/issues, etc. There is a solid bit of learning in the prompts, that the article uses.
Moving on from analyzing dish preparation videos to audio generation. The whole world has been mesmerized by NotebookLM and its ability to generate podcasts from content. What if you’d like to understand what it takes to generate a podcast with Flash 1.5 Pro? Here’s a code heavy blog post that shows the process.
If you’ve been building Generative AI applications over the year, grounding your Generative AI solutions in real-world data is the key. And if you are using Google Cloud Vertex AI, then what are the options available to you to ground your agents in enterprise data? Check out this blog post.
While on the topic of building a RAG solution, how about a blog post that highlights some of the best practices while building it out? Check out this blog post, that specifically explains to consider a transparent, automated evaluation framework, that helps to improve the quality. The post then eventually graduates into suggesting the Google Cloud’s generative AI evaluation service.
Agentic Applications are expected to be the area, where the AI momentum will shift to in 2025. Towards this, Google Cloud has announced an early access program called Google Agentspace, that primarily looks at tapping enterprise data, with the following 3 features:
- New ways to interact and engage with your enterprise data using NotebookLM
- Information discovery across the enterprise
- Expert agents to automate your business functions
Check out the blog post to learn more.
Vertex AI Search is one of the most powerful services in the Google Cloud AI Portfolio. It can search from billions of semantically similar or semantically related items. But how does it actually work? What are the real world use cases that customers have applied it to? Check out this detailed blog post.
Infrastructure
Google Cloud has announced its 41st cloud region in Querétaro, Mexico. Querétaro now joins Santiago, Chile, and São Paulo, Brazil as the 3rd region in Latin America. Check out the list of Google Cloud locations.
Trillium, the sixth-generation and most performant TPU to date, is now available to all Google Cloud customers. An interesting data shared was that Google used Trillium TPUs to train the new Gemini 2.0. Additionally, Google Cloud’s AI Hypercomputer has seen some new announcements too: key updates have happened to optimizations to the XLA compiler and popular frameworks such as JAX, PyTorch and TensorFlow to achieve leading price-performance at scale across AI training, tuning, and serving. Check out the blog post to see the across the board improvements to this architecture.
Databases
If you’d like to get all the updates in the world of Google Cloud Databases recently, then check out the single blog post.
Spanner saw some significant announcements throughout the year. A blog that captures “Spanner in 2024: A year of innovation” talks about how Spanner announced multiple “functional capabilities, pushed the envelope on price-performance, re-architected it for best-in-class reliability and security, and enhanced the developer experience”.
Database Center, which provides a fleet management solution to help manage your databases at scale has expanded its support for additional databases: Bigtable, Memorystore, and Firestore databases in preview. This is in addition to Cloud SQL, AlloyDB and Spanner that was announced in October 2024. Check out the blog post for details on the capabilities introduced for the newly added databases.
Google Cloud has announced a new set of integration capabilities for LlamaIndex, which is a popular open-source framework for Build AI Knowledge Assistants over your enterprise data. You would have guessed by now that for enterprise data that you have in AlloyDB for PostgreSQL and Cloud SQL for PostgreSQL, you can now take advantage of LlamaIndex integrations available for the same. Check out the blog post for details.
If you’ve gone through the last 2 years of Google Cloud, without taking a look at AlloyDB, you can check out this year-end post that provides an AlloyDB overview, its key benefits and an e-book that you can download.
Data Analytics
Cloud Pub/Sub is an often behind the scenes service that brings multiple Google Cloud services together. Specifically in 2024, Cloud Pub/Sub focused on streaming and particularly features that address three key data analytics patterns: Streaming ingestion, Streaming analytics, Stream sharing and export. Check out the blog post that provides more details on these updates in the post titled “Cloud Pub/Sub 2024 highlights: Native integrations, sharing and more”.
Customer Stories
Unexpected cloud costs can occur due to a variety of reasons. Some of these are misconfigurations, unexpected rise in traffic, lack of FinOps discipline and more. To address these situations, Google Cloud has a Cost Anomaly Detection as part of the Cost Management toolkit. Leading cyber security organization, Palo Alto Networks, went a step further. As the blog post states “it wanted a service that could identify anomalies based on labels, such as applications or products that span across Google Cloud projects, and provide more control over anomaly variables that are detected and alerted to its teams. Creating a consistent experience across its multicloud environments was also a priority.” Check out the blog post for details on this solution.
Management Tools
One of the most requested features by Organizations using Google Cloud and using Gemini, has been tracking the usage of Gemini across the organization. The natural place for these logs to appear is the Cloud Logging and Cloud Monitoring. This is now available. Currently in public preview, Cloud Logging records requests and responses between Gemini for Google Cloud and individual users, while Cloud Monitoring reports 1-day, 7-day, and 28-day Gemini for Google Cloud active users and response counts in aggregate. Check out the blog post on this feature.
Networking
The second installment of Network Performance Decoded, a series of whitepapers sharing best practices for network performance and benchmarking, has been released. This installment takes a look at Network Performance Limiters and how you can work around them. The paper also looks like TCP Round Trip Time (RTT) and Path MTU discovery, a process that helps prevent IP fragmentation.
Developers and Practitioners
Gemini Code Assist, an AI coding assistant, has announced support for latest version of the Gemini 2.0 models. The goal with the newer model is higher quality responses and lower latency. What is interesting to note in the blog post announcement is the growing ecosystem of tools around Gemini Code Assist, which should provide additional assistance across the SDLC and keep users in the IDE and thereby flow. The new tools feature in Gemini Code Assist includes Atlassian (Rovo), GitHub, GitLab, Google Docs, Sentry, and Snyk, with more to come.
Looking to build a customized search engine for an ecommerce marketplace using Spanner? A search engine that requires that you combine vector search, full-text search, and machine learning (ML) model reranking capabilities. Check out this blog post that highlights how Spanner makes this possible.
Contest Alert
If you are a baseball fan and a user of Google Cloud services with Gemini, you have a chance to develop an interesting solution, hit the ball out of the park and win some cool prizes at Google Cloud X MLB Hackathon.
You can combine these services with Major League Baseball (MLB) datasets to create something in the following 5 categories: Wild card — fan experience, Personalized fan highlights, Real-time “tool tips”, Generate Statcast data from old videos and Prospect prediction.
Check out the blog post and if you want to jump right to the contest page, go here: next2025challenge.devpost.com.
Learn about Google Cloud
If you have some time this holidays, you can catch up on 12 days of learning on Generative AI, made available by Google Cloud. And if you’ve not yet joined the Innovators community, please do so. You get to activate 35 monthly learning credits in Google Cloud Skills Boost at no cost and use these credits to access courses and labs.
A list of the first 6 courses is given below:
Stay in Touch
Have questions, comments, or other feedback on this newsletter? Please send Feedback.
If any of your peers are interested in receiving this newsletter, send them the Subscribe link.
