Google Cloud Platform Technology Nuggets — April 1–15, 2024 Edition

Romin Irani

Published in

Google Cloud - Community

12 min readApr 17, 2024

Welcome to the April 1–15, 2024 edition of Google Cloud Platform Technology Nuggets.

Please feel free to give feedback on this issue and share the subscription form with your peers.

Google Cloud Next 2024

Google Cloud Next 2024 was held from April 9–11, 2024 in Las Vegas and Generative AI was centrestage across the keynotes, sessions and more. It is difficult to put down all the blog posts that go into details of what was announced/demonstrated at the conference, but here are few links worth your time to catch up on the announcements:

Cloud Next ’24 Opening Keynote in under 14 minutes
Welcome to Google Cloud Next 24 : A text form of key announcements made in the Keynote.
All 218 things announced at Cloud NEXT ‘24
Day 1 Recap: AI Agents for everyone
Day 2 Recap: building AI Agents

The sessions are all available on demand at the Cloud NEXT ’24 site. The only requirement is that you will need to register to view the sessions. Check out the entire session library from the conference.

The other sections in this newsletter will highlight key announcements pertaining to those areas.

Infrastructure

Google Cloud’s Compute portfolio saw some key announcements vis-a-vis workload optimized infrastructure. This new introductions included:

C4 and N4, new general-purpose VMs powered by 5th Generation Intel Xeon processors
Upcoming preview of native bare-metal C3 machine types
X4, Compute Engine’s new memory-optimized instances, now in preview.
Z3, Google Cloud’s first storage-optimized VM family
Google Axion Processor, a new Arm-based CPU

Check out the blog post for more details.

Google Axion Processors, our first custom Arm®-based CPUs designed for the data center has been announced. Built using the Arm Neoverse™ V2 CPU, the Axion processors are expected to deliver significant performance improvements for general-purpose workloads like web and app servers, containerized microservices, open-source databases, in-memory caches, data analytics engines, media processing, CPU-based AI training and inferencing, and more. This compute option is likely to be available to customers later in the year. Check out the blog post for more details.

Two new Open Source offerings in the area of AI Inferencing have been announced: JetStream and MaxDiffusion. Check out the blog post.

A suggested read in this space is the key enhancements that Google Cloud has been doing in its AI Hypercomputer architecture space. Check it out.

Containers and Kubernetes

Here is a fantastic post that highlights how Container Services on Google Cloud are set to efficiently serve the needs of Gen AI applications in the future. The post nicely highlights to get started with Cloud Run for an easy AI starting point, GKE for training and inference and more. Check out how Google Cloud is positioning these services to continue serving key workloads via containers.

Looking to serve and deploy Gemma on GKE Standard as well as Autopilot clusters, check out a blog post that highlights the key enhancements that have been made to GKE to help you do that. The post includes Integration with Hugging Face, Kaggle and Vertex AI Model Garden, GKE notebook experience using Colab Enterprise and A cost-efficient, reliable and low-latency AI inference stack.

GKE Autopilot announced support for burstable workload support. As the blog post states, “Bursting allows your Pod to temporarily utilize resources outside of those resources that it requests and is billed for.” Another feature announced was Pods as small as 1/20th of a vCPU can be used now. Additionally, you can create any size of Pod you like between the minimum to the maximum size, for example, 59m CPU, 302m, 808m, 7682m etc. These features when combined together is a great way to run high-density workloads. Check out the blog post for more details.

Networking

A big announcement was Cloud Service Mesh, a fully managed service mesh across all Google Cloud platform types. This is a single offering that combines Traffic Director’s control plane and Google’s open-source Istio-based service mesh, Anthos Service Mesh. Check out the blog post that highlights what customers get from this new offering and the benefits.

Google Cloud’s next-generation cloud firewall offering is now available in GA. The product Cloud Firewall Plus, now called Cloud NGFW (Next Gen Firewall) is available in 3 tiers: Essentials, Standard and Enterprise. Check out the blog post for more details.

If you’d like to get all the details on Whats New in Networking announced at Cloud NEXT 24, check out this blog post.

Identity and Security

Security saw some key updates at Google Cloud NEXT ’24. Gemini in Security Operations has a new assisted investigation feature, generally available at the end of this month, which will guide analysts through their workflow wherever they are in Chronicle Enterprise and Chronicle Enterprise Plus. Gemini recommends actions based on the context of an investigation, and can run searches and create detection rules to improve response times. You can also ask Gemini for the latest threat intelligence from Mandiant directly in-line — including any indicators of compromise found in their environment — and Gemini will navigate users to the most relevant pages in the integrated platform for deeper investigation.

This is just one of several updates in the Security space. Check out the blog post for more security related announcements.

Machine Learning

Gemini is now right across Google Cloud. Whether you are a developer, security analyst, data analyst, operator, etc — you are bound to use Gemini now across various services. The names can get a bit confusing and hence it is important that you read this article first to get the product names right and what Gemini does to help work with specific areas. Check out the post that talks more about the diagram that you see below.

Once you are done with the above post, do check out a post that highlights updates to Gemini, Imagen, Gemma and MLOps on Vertex AI. The updates include Gemini Pro 1.5 now available in Public Preview, Imagen’s new text-to-live image capabilities and more.

When it comes to using Generative AI in Enterprise applications, grounding these systems to the truth is essential and an absolute requirement. Google Cloud defined “enterprise truth” as the approach to grounding a foundation model in web information; enterprise data like databases and data warehouses; enterprise applications like ERP, CRM, and HR systems; and other sources of relevant information. And how does it propose to do that? Via a preview of Ground with Google Search in Vertex AI. Check out the post for more details.

Vertex AI Search and Conversation products, along with other developer tools is all coming together under one umbrella: Vertex AI Agent Builder. The key objective is to help build our AI Agents that are grounded in factuality and uses key features like Vertex AI Extensions, function calling and data connectors. Check out the blog post for more details.

If you have been tracking Vertex AI Text Embeddings, at Cloud NEXT ’24, two new text embeddings were announced:

English only: text-embedding-preview-0409
Multilingual: text-multilingual-embedding-preview-0409

Storage and Data Transfer

Before Cloud NEXT ’24 where several storage innovations were announced, there were a couple of Storage updates:

You can now leverage the power of Google Cloud tags, including inheritance, to easily configure backup policies for Compute Engine VMs, ensuring consistent protection of your dynamic cloud environments. Check out the post.
You can now look at meeting data retention compliance via the new Object retention lock. This feature helps you set and lock retention configurations on Cloud Storage objects, with a “retain until time.” This means that an object with an object retention lock can not be deleted or replaced until the retain until time has passed. Check out the post.

At Cloud NEXT 24, with GenAI clearly dominating the mindspace, optimized storage solutions and features were announced to address the challenge of decreasing model load, training, and inference times while maximizing accelerator utilization. These included Cloud Storage FUSE with file caching, ParallelStore, Hyperdisk ML and more. Check out the blog post for more announcements made.

Databases

Gemini in Databases was a key announcement in the area of Databases at NEXT. This meant AI-powered assistance to simplify all aspects of the database journey including developing, monitoring, optimizing, securing and migrating database-driven applications, vector support across more of the databases and more. Check out the blog post for more details.

Check out this additional blog post that dives into Gemini for Databases and gives a preview of the AI assisted features in Database Studio.

Expanding on the above, AlloyDB for PostgreSQL got natural language support. This feature is key in helping developers integrate real-time operational data into generative AI applications. AlloyDB also introduced parameterized secure views, a new kind of database view that locks down access to end-user data at the database level to help you protect against prompt injection attacks.

Also announced is the new ScaNN index for AlloyDB, bringing 12 years of Google research and innovation in approximate nearest neighbor algorithms to AlloyDB. This index is said to “deliver up to 4x faster vector queries, up to 8x faster index build times and typically a 3–4x smaller memory footprint than the HNSW index in standard PostgreSQL”. At the moment, it is available in technology preview in AlloyDB Omni, and will become available in the AlloyDB for PostgreSQL managed service in Google Cloud later. Check out the blog post for more details.

I love the statement “How do you store the entire Internet?” in the following blog post, that highlights the goal that the team set out with, designed BigTable for that and now 20 years later, continue to boost BigTable with features that are expected to bring it to a more widestream set of users.

Looking to do migrations of SQL Server to the Cloud SQL for SQL Server managed service? A preview of support for SQL Server migrations to Cloud SQL for SQL Server in Database Migration Service is now available. DMS is a fully managed serverless cloud service that performs database migrations with minimal downtime. Check out a blog post how Database Migration Service works.

Speaking of migrations, Gemini in Databases and specifically when it comes to migrations can help out in multiple ways. How about explainability when it comes to understanding existing queries that you need to migrate, schema conversions and more. Check out an interesting blog post that highlights how you can use Gemini to help migrate Oracle to Cloud SQL for PostgreSQL on Google Cloud.

Private Service Connect is a capability of Google Cloud networking that allows consumers to access managed services privately from inside their VPC network. Private Service Connect is now fully integrated with Cloud SQL, Google Cloud’s fully managed database service for PostgreSQL, MySQL, and SQL Server. Check out the blog post for details on how to get started, configuring Private Service Connect and deployment architectures.

Memorystore for Redis Cluster saw some major announcements at Cloud NEXT:

Public preview of data persistence for both RDB (Redis Database) and AOF (Append Only File)
General availability of new nodes types of 1.4 GB, 6.5 GB and 58 GB
General availability of ultra-fast vector search on Memorystore for Redis
Public preview of new configuration options

Check out the blog post for more details.

Data Analytics

If you are looking at scanning just the key announcements, check out this post. Now let’s dive into some of those announcements.

BigQuery is now the single, AI-ready data analytics platform. A single product that helps you manage structured data in BigQuery tables, unstructured data like images, audience and documents, and streaming workloads, all with the best price-performance. Dive into this post to understand how.

Duet AI in BigQuery is now Gemini in BigQuery. Key assistance is now available in AI augmented data preparation that helps users to cleanse and wrangle their data. Another interesting feature is the new semantic search capabilities to help you pinpoint the most relevant tables for your tasks. Leveraging the metadata and profiling information of these tables from Dataplex, Gemini in BigQuery surfaces relevant, executable queries that you can run with just one click. Check out the post for more details.

The new BigQuery data canvas provides a reimagined natural language-based experience for data exploration, curation, wrangling, analysis, and visualization, allowing you to explore and scaffold your data journeys in a graphical workflow that mirrors your mental model. Check out this post.

The deep integration now of Gemini models into Looker is also going to open up multiple possibilities in the realm of Looker as a BI Platform. Check out this post.

A couple of other posts that are interesting and which give a glimpse into how BigQuery and GenAI capabilities have merged are:

Developers and Practitioners

Duet AI for Developers got rebranded as Gemini Code Assist, which is set to supercharge your development workflow. There were key demos provided at Cloud NEXT Developer Keynote that showed extended capabilities that include full codebase awareness, increase in local context, code transformation support (refactoring, etc), connecting to existing source repositories and various partner integrations. Check out the blog post that highlights each of these areas.

Firestore saw some key announcements at NEXT 24. These include:

Use Gemini Code Assist in your favorite Integrated Development Environment (IDE) to use natural language to define your Firestore data models and write queries.
Firestore now has built-in support for vector search using exact nearest neighbors, the ability to automatically generate vector embeddings using popular embedding models via a turn-key extension, and integrations with popular generative AI libraries such as LangChain and LlamaIndex. Check out a detailed article on using Firestore vector similarity search.
Firestore now supports Customer Managed Encryption Keys (CMEK) in preview, which allows you to encrypt data stored at-rest using your own specified encryption key.
Retain daily backups using Firestore’s Scheduled Backup feature for up to 98 days, up from seven days.

An interesting new service has been announced at Cloud NEXT 24 : App Hub. Think of one or more applications that you have deployed on Google Cloud. It is a challenge to understand visualizing that application in terms of not just the cloud resources that it uses but dependencies on other services, across project services and more. App Hub is targetted to address that by introducing abstractions in the form of Applications, Workloads and Services. It is able to injest automatically key Google Cloud services that Applications would use and then build out a dependency graph that is kept updated all the time. The current resources that it supports are various Load Balancing services (Services) and Compute Engine MIGs (Workloads). As the service grows to support GKE and Cloud Run, it could get very useful. Check out the blog post.

If you are using Apigee, Gemini Code Assist is making its way into the product suite to help ease the task of developing API specifications, API integrations and more. For example, when it comes to building out APIs, you can build out an API Specification using the Gemini Code Assist integration inside of the Cloud Code VS Code extension. These specifications can then be published to the API Hub. Not just that but Gemini offers step-by-step guidance for adding new policy configurations while creating an API proxy. Lastly, Gemini also provides explanations for your existing configurations, reducing the learning curve during updates and maintenance. Check out the post for more details.

Learn Google Cloud

When it comes to service discovery and DNS resolution in your GKE clusters, you have a choice with kube-dns, Cloud DNS, etc. Deep dive into this options via this informative blog post.