Tibor Kiss

Google Cloud Next is heading to Las Vegas in 2024 from April 9th to 11th, and I’m excited to announce that registration opens today. With a new location, fresh programming, hundreds of partners, and the latest in AI, analytics, security, infrastructure, collaboration and more, Google Cloud Next ‘24 is set to be our biggest event yet.

Early registration at the Google Cloud Next site is now open at a discounted rate of $999 USD* (that’s 50% off the full ticket price of $1,999 USD), and includes one year of Innovators Plus (limited quantity), so there’s no better time to register. Last year’s event with over 20,000 people sold out quickly — check out our 2 min recap video.

There has never been a more exciting time to learn more about cloud technology, and staying up-to-date on the latest in AI innovation from Google experts, partners, and your peers is critical for anyone looking to keep pace in this rapidly-evolving space. Throughout the conference, you’ll have opportunities to connect directly with Google engineers and experts, who can help you solve problems and brainstorm innovative solutions 1:1.

At Cloud Next ’24, we have learning opportunities for everyone.

For developers, there will be hundreds of learning sessions and an Innovators Hive with “how-to” content and demos, in addition to hands-on training on gen AI and other cutting-edge cloud technologies.For leaders we will highlight dozens of gen AI use cases in every function and industry and share details on how to move quickly from proof-of-concept to production as well as expected outcomes.For startups, we have a curated experience that includes startup sessions, networking opportunities, and the bustling Startup Lounge, where you can learn about how Google Cloud can help supercharge your startup’s growth.For everyone, we will have a huge show floor featuring the latest and greatest from our Google Cloud partner ecosystem as well as a customer innovation showcase.

There will be dedicated sessions for all levels across all roles, including application developers, architects and IT professionals, data and database engineers, data scientists and data analysts, DevOps/SREs/ITOps/platform engineers, IT managers and business leaders, productivity/collaboration app makers, security professionals, startup leaders/founders – there’s truly something for everyone!

Register today at our dedicated Google Cloud Next ‘24 site.

*The $999 USD early bird price is valid through 11:59 PM PT on Wednesday, January 31, 2024 or until it’s sold out.

Everyday, more and more small businesses and startups are leveraging generative AI for their internal and customer-facing needs, with over 70% of gen AI’s billion dollar-unicorns using Google Cloud. As the technology continues moving at a rapid pace, with almost daily announcements of products and features, businesses are discovering new use cases and ways to leverage gen AI tools.

In our monthly gen AI workshops, we’re meeting small businesses to provide them with a technical overview of our key gen AI products, AI techniques, and best practices. Our workshops are organized and led by a team of Specialist Customer Engineers, Monique Sullivan, Bruk Zewdie, Isaac Attuah, and myself. They ignite imagination, as together, we brainstorm and ideate with our customers to blend their creativity with Google Cloud’s hands-on experience deploying large models. Our workshop provides a collaborative environment to unearth innovative applications and use cases. Working with our team, we explore industry use cases, confront and assess common challenges, and identify successful techniques in deployment strategies.

If you are a startup or small business looking for some inspiration on how to use gen AI, this article will explore some of the prominent themes and trends from our most recent workshops—and explain how you can participate in a future edition.

Popular use cases

Customers across a range of industries are using the power of generative AI, from doctors helping patients, to businesses empowering employees and customers, to individuals bridging knowledge gaps or enhancing visual creativity. Below are some of the industry use cases that have sparked excitement in our recent sessions.

Healthcare:

Carry medical assistance in your pocket—Help doctors focus on patients rather than paperwork, with an AI physician’s assistant always at the ready.Get a bespoke workout routine via an AI personal trainer that tailors its advice from a library of fitness videos and regimens.Enable precision quality control – evaluating healthcare provider performance using statistical analysis of clinical data.

Employee Support:

Instill company wisdom with an HR chatbot answering policy questions 24/7, guiding employees with institutional knowledge.Automate headshot quality by leveraging AI to ensure polished employee profile pics, optimized with studio-quality lighting and touch-ups.Quickly find answers with a large language model(LLM)-generated knowledge base—a personalized, up-to-date employee handbook.

Customer Support:

Deliver instant, personalized support with chatbots that engage customers using natural conversations and contextual understanding.Remember each customer via support bots that tap into full histories and preferences for tailored, seamless issue resolution.Make better decisions with accurate forecasting at your fingertips, powered by an AI assistant that predicts outcomes using past data.

Visual-based:

Accelerate game development by leveraging AI to automatically generate graphics and visual assets.Gain actionable insights by analyzing videos with AI to identify trends, movements, and anomalies.Automate custom illustrations, using AI to create original drawings for children’s books to spark young imaginations.Condense information into dynamic videos, transforming documents into shareable lessons with AI-generated animations.

Bridging Gaps:

Making data accessible—automatically translate complex financial reports into plain language for any audience.Drive personalized recommendations, using AI to analyze assessment data and propose customized teaching strategies.

Successful prompting techniques:

In addition to use cases, workshop participants are also eager to understand how to more effectively interact with generative models— that is, how to prompt them.

Prompting is the art of guiding an LLM with your input so it provides your desired response. Throughout our workshops, our attendees share and unlock effective techniques like:

Establish the LLM’s expertise and role upfront in your prompt so it understands its persona.Have the LLM restate prompts to ensure it’s correctly processed your input.Use structured instructions for clarity: “summarize the input: [text.]”Ask the LLM to craft specific prompts tailored for your needs.Streamline prompts like Google searches – no fluff.Organize the prompt with lists or dictionaries.Insert “variables” into the prompt, allowing users to customize the prompt by just changing the variable words.Prompting is a skill that mixes your personal creativity with your industry knowledge – try multiple approaches to get different results.

Join our next Generative AI workshop

See gen AI in action at our next Generative AI Workshop. Preview tools and services for startups and small businesses with our specialist team. Over two hours, you and your team will have the opportunity to see a deep dive and demos on the upcoming Google Cloud gen AI product portfolio. You’ll leave equipped with strategies to implement cutting-edge gen AI solutions that give your business a competitive advantage. The workshop will also give you the opportunity to receive free credits to labs with a sandbox environment so you can experiment and get started with your generative AI journey.

The future is generative – join us to seize it and propel your startup forward.

To learn more about how Google Cloud can help your startup, visit our pageherefor more information about our program, where startups can get up to $200,000 USD (up to $350,000 USD for AI startups) in Google Cloud and Firebase cost coverage over 2 years, as well as technical training, business support, and Google-wide offers.

Geographical redundancy is one of the keys to designing a resilient data lake architecture in the cloud. Some of the use cases for customers to replicate data geographically are to provide for low-latency reads (where data is closer to end users), comply with regulatory requirements, colocate data with other services, and maintain data redundancy for mission-critical apps.

BigQuery already stores copies of your data in two different Google Cloud zones within a dataset region. In all regions, replication between zones uses synchronous dual writes. This ensures in the event of either a soft (power failure, network partition) or hard (flood, earthquake, hurricane) zonal failure, no data loss is expected, and you will be back up and running almost immediately.

We are excited to take this a step further with the preview of cross-region dataset replication, which allows you to easily replicate any dataset, including ongoing changes, across cloud regions. In addition to ongoing replication use cases, you can use cross-region replication to migrate BigQuery datasets from one source region to another destination region.

How does it work?

BigQuery provides a primary and secondary configuration for replication across regions:

Primary region: When you create a dataset, BigQuery designates the selected region as the location of the primary replica.Secondary region: When you add a dataset replica in a selected region, BigQuery designates this as a secondary replica. The secondary region could be a region of your choice. You can have more than one secondary replica.

The primary replica is writeable, and the secondary replica is read-only. Writes to the primary replica are asynchronously replicated to the secondary replica. Within each region, the data is stored redundantly in two zones. Network traffic never leaves the Google Cloud network.

While replicas are in different regions, they do not have different names. This means that your queries do not need to change when referencing a replica in a different region.

The following diagram shows the replication that occurs when a dataset is replicated:

Replication in action

The following workflow shows how you can set up replication for your BigQuery datasets.

Create a replica for a given dataset

To replicate a dataset, use the ALTER SCHEMA ADD REPLICA DDL statement.

You can add a single replica to any dataset within each region or multi-region. After you add a replica, it takes time for the initial copy operation to complete. You can still run queries referencing the primary replica while the data is being replicated, with no reduction in query processing capacity.

To confirm the status that the secondary replica has successfully been created, you can query the creation_complete column in the INFORMATION_SCHEMA.SCHEMATA_REPLICAS view.

Query the secondary replica

Once initial creation is complete, you can run read-only queries against a secondary replica. To do so, set the job location to the secondary region in query settings or the BigQuery API. If you do not specify a location, BigQuery automatically routes your queries to the location of the primary replica.

If you are using BigQuery’s capacity reservations, you will need to have a reservation in the location of the secondary replica. Otherwise, your queries will use BigQuery’s on-demand processing model.

Promote the secondary replica as primary

To promote a replica to be the primary replica, use the ALTER SCHEMA SET OPTIONS DDL statement and set the primary_replica option. You must explicitly set the job location to the secondary region in query settings.

After a few seconds, the secondary replica becomes primary, and you can run both read and write operations in the new location. Similarly, the primary replica becomes secondary and only supports read operations.

Remove a dataset replica

To remove a replica and stop replicating the dataset, use the ALTER SCHEMA DROP REPLICA DDL statement. If you are using replication for migration from one region to another region, delete the replica after promoting the secondary to primary. This step is not required, but is useful if you don’t need a dataset replica beyond your migration needs.

Getting started

We are super excited to make the preview for cross-region replication available for BigQuery, which will allow you to enhance your geo-redundancy and support region migration use cases. Looking ahead, we will include a console-based user interface for configuring and managing replicas. We will also offer a cross-region disaster recovery (DR) feature that extends cross-region replication to protect your workloads in the rare case of a total regional outage. You can also learn more about BigQuery and cross-region replication in the BigQuery cross-region dataset replication QuickStart.

Since we launched Memorystore for Redis Cluster in preview, customers across a variety of industries including banking, retail, ads, manufacturing, and social media have taken advantage of the performance and scalability of Memorystore for Redis Cluster.

Today we’re thrilled to announce the general availability (GA) of Memorystore for Redis Cluster, providing a 99.99% SLA with replicas enabled. With Memorystore for Redis Cluster, you get a fully-managed and fully open-source software (OSS) compatible Redis Cluster offering with zero downtime scaling (in or out), providing up to 60 times more throughput than Memorystore for Redis, with microseconds latency. Memorystore for Redis Cluster intelligently places primaries and replica nodes across availability zones and manages automatic failovers to maximize availability and reduce complexity.

Performance sensitive customers like PLAID, Wright Flyer Studios, and Reddit are excited about this GA and the 99.99% SLA.

“Memorystore for Redis Cluster provides us the scalability and manageability required for our workloads, allowing us to use Memorystore in far more meaningful ways. The OSS cluster client compatibility also provides an easier migration path from existing self-managed Redis Clusters. Now that it is GA with a 99.99% SLA, we look forward to adopting it for our caching needs on Google Cloud!” – Stanley Feng, Senior Engineering Manager at Reddit

Performance and scaling

Memorystore for Redis Cluster takes Memorystore to the next level of scale and performance, supporting 10 times more data and providing up to 60 times more throughput than Memorystore for Redis, with microseconds latency. Memorystore’s vetted, production-ready infrastructure provides a pre-configured and optimal Redis platform to supercharge your performance. With zero downtime scaling, you can start small and grow over time, adding only the capacity required based on workload demands, without taking downtime and interrupting your production applications. Unlike client-sharded caching architectures, which are prone to data loss during scaling operations, the simple scaling APIs of Memorystore for Redis Cluster let you grow or shrink your clusters and optimize costs without data loss. The direct access (proxy-less) architecture of Memorystore for Redis Cluster ensures that throughput scales linearly as nodes are added to the cluster, providing ultra-low and predictable latencies and avoiding the cost and latency overhead of a proxy-based architecture.

Applibot, a CyberAgent company, produces and operates smartphone games and services, and depends heavily on Redis and its ultra-fast performance:

“After relying on Memorystore, we’re excited for the launch of Memorystore for Redis Cluster, which will dramatically improve performance with its ultra low and predictable latencies. We’re also very excited to take advantage of Memorystore for Redis Cluster’s easy-to-use APIs which make scaling clusters in or out simple, fast, and non-disruptive, enabling us to dynamically size the clusters based on workload demands.” – Naoto Ito, Lead of Backend Engineering at Applibot

High availability

Memorystore for Redis Cluster is built with a robust control plane that provides high availability with data replication across multiple fault-domains (zones), automatic failovers, and intelligent node placement for both primary and replica nodes. Memorystore for Redis Cluster can handle both node and zonal outages by efficiently failing over to the replica nodes, promoting them to become primaries, and automatically repairing failed nodes with seamless orchestration. Memorystore for Redis Cluster’s durable architecture is designed for reliability, as Redis clients directly access the shards (primary and replica nodes). This design avoids the risk of a single point of failure (as each shard is designed to fail independently), which is inherent in proxy-based architectures.

In addition, Memorystore for Redis Cluster’s control plane utilizes Create-Before-Destroy and graceful shutdown maintenance strategies to eliminate downtime or workload interruptions. Instead of upgrading active cluster nodes in-place, Memorystore first creates an entirely new replica with the new software (at no extra charges to you), then coordinates a lossless failover from the old node to the new node, and lastly removes the old node gracefully in a gradual process that ensures minimal-to-no impact to your application.

NTT DOCOMO, Japan’s largest telecommunications company, is thrilled with the general availability of Memorystore for Redis Cluster:

“Memorystore for Redis Cluster’s resilient design and 99.99% SLA gives us confidence that our clusters are durable and resilient. Memorystore’s sophisticated control plane and maintenance orchestration ensures minimal impact to our application during maintenance and we’re very excited to utilize this managed offering so we can focus our efforts on creating value for our customers.” – Masatoshi Kato, Manager, Service Design Department, NTT DOCOMO

Integrated and automated with Google Cloud

Memorystore for Redis Cluster is built with Private Service Connect (PSC) for private and secure connectivity by default. With PSC, we simplified the provisioning experience so you can easily configure private networking with an automated single-step process, avoid security issues of bidirectional access, and not be limited by the quota issues posed by Virtual Private Cloud (VPC) peering-based implementations. We also simplified cluster endpoint management by only requiring two IP addresses for any sized cluster (even for 250 Redis nodes!), thereby addressing IP exhaustion and contiguous expansion problems. With PSC, client applications can easily access the cluster from any region and there are advanced security controls available, like separation of permissions for Network and Redis admin and control of the IP address space used for cluster endpoints.

Memorystore for Redis Cluster comes with built-in integration with Google Identity and Access Management (IAM) so you can easily manage and control access to your clusters, especially for a microservice based client architecture. We also provide out of the box Audit Logging, in-transit encryption with TLS, and integration with Cloud Monitoring, so your cluster metrics are accessible and actionable.

Fully-managed vs do-it-yourself

For many customers, it’s the fully-managed nature of the Memorystore for Redis Cluster that entices them to migrate. You can offload the tedious and unrewarding work of tuning Redis performance on Compute Engine VMs, orchestrating complex cluster topologies, optimizing networking implementations, managing maintenance, and striving for self-managed high availability in the face of outages and upgrades. Customers can rely on Google to continually invest in building and supporting new Redis features and delivering them to customers with non-disruptive maintenance operations. Now, you can focus on building value for your organizations instead of focusing on managing Redis Clusters. To make adoption even more enticing, Memorystore Committed Use Discounts (CUDs) are already available for Memorystore for Redis Cluster and are fungible with Memorystore for Redis as well as Memorystore for Memcached, enabling savings of 40% with a three-year CUD and 20% with a one-year CUD.

Waze, a free navigation and driving app active in more than 185 countries is adopting Memorystore for Redis Cluster to take advantage of the fully-managed offering and supercharge their applications’ performance:

“Waze is excited to use Memorystore for Redis Cluster as our primary cache solution, taking advantage of the 99.99% SLA, zero-downtime scalability, and flexible Redis data types. The ultra-fast performance has been instrumental in helping us scale our platform and provide a better experience for our users.” – Yuval Kamran, DevOps Engineer

Getting started

At this point, you may be wondering how to get started or how much effort it would require to migrate to the Memorystore for Redis Cluster. Good news! We’ve already published a zero-downtime migration blog using OSS RIOT with code snippets and a detailed step-by-step walkthrough so you can quickly and easily migrate your Redis implementations from any source, including self-managed or third-party offerings without interrupting your applications.

You can easily get started today by heading over to the Cloud console and creating a cluster, and then scaling in or out with just a few clicks. Please let us know if you have any feedback by reaching out to us at cloud-memorystore-pm@google.com.

This year, Google has seen an increase in the number of vulnerabilities impacting central processing units (CPU) across hardware systems. Two of the most notable of these vulnerabilities were disclosed in August, when Google researchers discovered Downfall (CVE-2022-40982) and Zenbleed (CVE-2023-20593), affecting Intel and AMD CPUs, respectively.

This trend proves only to be intensifying as time goes on. Left unmitigated, these types of vulnerabilities can impact billions of personal and cloud computers.

Today, we’re detailing the findings of Reptar (CVE-2023-23583), a new CPU vulnerability that impacts several Intel desktop, mobile, and server CPUs. Google’s Information Security Engineering team reported the vulnerability to Intel, who disclosed the vulnerability today. Thanks to the thoughtful collaboration between Google, Intel, and industry partners, mitigations have been rolled out, and Googlers and our customers are protected.

How Google found and responded to Reptar

A Google security researcher identified a vulnerability related to how redundant prefixes are interpreted by the CPU which leads to bypassing the CPU’s security boundaries if exploited successfully. Prefixes allow you to change how instructions behave by enabling or disabling features. The full rules are complicated, but in general, if you use a prefix that doesn’t make sense or conflicts with other prefixes, we call those redundant. Usually, redundant prefixes are ignored.

The impact of this vulnerability is demonstrated when exploited by an attacker in a multi-tenant virtualized environment, as the exploit on a guest machine causes the host machine to crash resulting in a Denial of Service to other guest machines running on the same host. Additionally, the vulnerability could potentially lead to information disclosure or privilege escalation.

You can read more technical details about the vulnerability at our researcher’s blog.

Our security teams were able to identify this vulnerability and responsibly disclose it to Intel. Google worked with industry partners to identify and test a successful mitigation so all users are protected from this risk in a timely manner. In particular, Google’s response team ensured a successful rollout of the mitigation to our systems before it posed a risk to our customers, mainly Google Cloud and ChromeOS customers.

Google’s commitment to collaboration and hardware security

As Reptar, Zenbleed, and Downfall suggest, computing hardware and processors remain susceptible to these types of vulnerabilities. This trend will only continue as hardware becomes increasingly complex. This is why Google continues to invest heavily in CPU and vulnerability research. Work like this, done in close collaboration with our industry partners, allows us to keep users safe and is critical to finding and mitigating vulnerabilities before they can be exploited.

We look forward to continuing this proactive cybersecurity work, and encourage others to join us on this journey to create a more secure and resilient technology ecosystem.

Savimbo was founded to facilitate a better way to help stop deforestation of the Amazon rainforest by directly paying small farmers and indigenous groups to be good stewards of the land. Smallfarmers control 80% of the world’s tropical forests, guarding 70% of its carbon stores. Indigenous groups are 5% of the world’s population guarding 80% of its biodiversity. Our goal is to employ one billion small farmers within ten years to support our mission.

Our company, a B Corporation, was started by and for indigenous small farmers in the Colombian Amazon who have responsibly conserved their privately held lands for generations. To support these efforts, we take advantage of the latest technologies including Google Cloud so that small farmers can better preserve and replant the rainforest, retain biodiversity, and protect wildlife, while earning a fair income.

In addition to working with small farmers, we provide a product that large global corporations need: premium-quality and high-value biodiversity, water, tree, and carbon credits that are audited and reliable. All of these things come together to further corporate environmental, social, and governance (ESG) commitments.

Removing intermediaries with direct payments to small farmers

Our fair-trade biodiversity credits are a win-win for large corporations and digitally under-represented indigenous peoples alike. We eliminate the middlemen and complexity for small farmers. Standardizing on the best technologies makes our streamlined approach possible, both managing the biodiversity credits, as well as handling small farmer land enrollments, direct-to-farmer payments, and tracking and certification for reforestation and wildlife preservation.

For example, we will be usingFirebase Realtime Database to store tree-tracking data we gather from our small farmers and other sources. Changes to the data, such as when a new native tree is planted, are logged on smartphones, even while offline in rural areas of Putumayo, Colombia. Realtime Database SDKs sync the data automatically when the mobile phones connect to the internet later.

It needs to be easy for indigenous peoples and local communities to participate directly in our programs, eliminating the intermediary cost burdens that most other programs entail. That means that the methods need to be straightforward, with any complex calculations automated in code. We useGoogle Earth Engine for that, including access to satellite imagery and geospatial data that inform our analytics for tracking progress.

WithBigQuery andLooker Studio we will use AI and ML for predictive analytics, cultivate insights using real-time data, and prepare visualizations of our findings. This helps us continue to improve our approach to promoting biodiversity richness and offer a premium product that improves protecting the Earth.

Empowering self-sufficiency in ecological silviculture

A big part of Savimbo’s mandate is to train local talent with trustless verification methods instead of relying on outsiders to come in and manage projects. We have a core group of very bright, talented young people. Some of our team members who are doing Google Earth coding today were pig farming about six months ago.

We also have a consulting network of hundreds of people in 27 countries that speak 15 languages to advise Savimbo, many of them working pro bono. These include the five indigenous taitas (local medicine doctors) that were the genesis of our company, a Nobel Peace Prize winner, and a rocket scientist. Getting these diverse perspectives have allowed us to continue to be thoughtful as we build out our company.

Our three founders are Jhony, an animal rights activist and jaguar expert who has worked with the World Wildlife Foundation and in conservation for two decades. Drea has a background in science, medicine, and technology, plus experience in the field as a ‘hotshot’ wildland firefighter for the US Forest Service and for the British Columbia Ministry of Forests in Canada. Our third co-founder, Fernando Lezama, has been an indigenous rights activist and taita for 30 years.

We now have leaders from eighteen indigenous groups, 300 small farming families, and several governmental and non-governmental agencies partnering with us to protect 16,500 hectares so far. And we’re exploring expansion soon to Ecuador and Mexico.

Launching our company with support from theGoogle for Startups Cloud Program was really helpful. We received cloud credits that gave us a jumpstart on building our infrastructure. We also just joined theGoogle Startups for Sustainable Development program and are very excited about the mentorship, funding connections, and technology support it provides. We appreciate that Google Cloud understands our impact metrics and is on board with the idea of doing good while running a successful business, and we appreciate working with the cleanest cloud in the industry.

We hope to be a role model for planetary consciousness. If we are going to address the climate challenges in front of us, we have to think and collaborate beyond national borders and work together to be smarter about resources on a global level.

If you want to learn more about how Google Cloud can help your startup, visit our pagehereto get more information about our program, andsign up for our communicationsto get a look at our community activities, digital events, special offers, and more.