------------------------------
Title: Google Cloud Platform Technology Nuggets — February 16–29, 2024 Edition
[content]
Google Cloud Platform Technology Nuggets — February 16–29, 2024 EditionWelcome to the February 16–29, 2024 edition of Google Cloud Technology Nuggets.Please feel free to give feedback on this issue and share the subscription form with your peers.Google Cloud NEXT ’24Google Cloud NEXT ’24 is just around the corner. The event will be held from April 9–11, 2024 in Las Vegas and there are likely to be some cool announcements. In the meanwhile, check out the entire session catalog.Fifth “epoch” of computingThis is an excellent article that you should read that goes through various epochs or transitions in the last 50+ years. Each of these epochs have been transformative and the article argues that we are currently in the fifth epoch of computing, one that is data-centric, declarative, outcome-oriented, software-defined, and centered on proactively bringing insights to people. This epoch is supposed to be the most transformative and brings with it key challenges too. Check out the article.Infrastructure and Modernization of environmentsIf you have been using Compute Engine, you know that you can start and stop the Compute Engine instances. However this feature was not available if you used Managed Instance Groups (MIGs). This has been addressed now with the release of the standby pool feature in MIGs. A standby pool can be configured to contain the target size of the suspended and stopped VMs. This pool can be managed manually or via an automated fashion when the MIGs size is increased. Check out the article for more details and how you can start using this cost-saving feature today.Mainframes conjure up a very different era of computing. But for its time, organizations who used the technology setup a centralized way of controlling the access, security, operations and setup layers of processes since resources were scarce. With these organizations undergoing a transformation and embarking on a mainframe modernization journey, they have to adopt the ways of the cloud, especially when using a Cloud provider that provides managed services, where security is a shared responsibility, where application errors are only within the realm of the customer to understand and more. Check out this article that highlights why mainframe modernization is not just technology transformation but also of people and processes.And continuing on the theme of modernization, its interesting to see how Contact Center Technology completely transformed the IVR systems of old. The earliest adopters of this were the Telecom providers that integrated their systems with the modern Contact Center Software. But with the advent of GenAI, this space is again ripe for a transformation of a different kind. One that will see more intelligent agents that allow a more natural interaction, grounding of the models with the Organizations knowledge base via RAG techniques and more. Check out how contact centers are being transformed with Generative AI.API ManagementApigee and GKE represent the best of services out there in Google Cloud. But often, you need to connect the API layer to your backend services running on GKE in a secure fashion. For security reasons, you do not want to expose the GKE services publicly plus setting up private and secure connections is an administrative overhead. Enter Private Service Connect (PSC) and GKE Gateway to help address that. Read this article for more details on these services and how to configure a secure and private connection that bridges your API layer to the GKE services.CustomersSymphony provides solutions that manage collaborations across worldwide financial institutions and typically this would mean millions of messages across their platform on a daily basis. Their original solution was based on HBase and that turned quickly into a signficant overhead in terms of running the clusters and managing them. This article covers their move to Bigtable, the reasons for choosing it and how they planned the migration. The result is significant cost savings and operational efficiencies at scale. Check out the article for more details.Identity and SecurityStill using Service Account Keys at scale? If yes, security findings continue to report that one of the primary causes of security breaches are related to Service Account Keys. As an organization, you have to step up the game on preventing service account key generation and there are ways to address this in a systematic way with help from specific features in Google Cloud. You can look at Organization policies that prevent Service Account key creation and also detect/alert on instances of Org Policy changes and Service Account key creation. Check out this article on how to do that.Continuing on the topic of preventive measures that need to be taken to avoid issues in the future, the same applies to Vertex AI services. With organizations taking on a more proactive role for using AI services in their applications, you need to have Organization Policies in place to govern the usage of Vertex AI services. Check out this article for more details, that also highlights how Security Command Center Premium can help detect instances of these activities.In the recent blog, Palo Alto Networks highlights several techniques that exploit misconfigurations which could allow a malicious actor to move laterally in cloud environments. The interesting thing about this blog is how bad actors can potentially use tools that are essential for administrating the environments and the machines in them. For e.g. you would think that taking backups, SSH Keys and access to serial console are essential utilities and while that is true, seldom does one think of how bad actors could utilize them to compromise your cloud environment. Check out this article that gives 3 key examples with steps on how the attackers would go about compromising the systems and what you can do in Google Cloud to prevent that. Essential read for anyone managing Google Cloud environments.There are couple of CISO reports in this edition. The first report highlights the threats that Google Cloud continues to see and will likely be the trend for 2024. Its interesting to note that one of the biggest threats emerges from the organization’s own security hygiene and misconfigurations. So something that the organizations should pay more attention to.The second report focuses on the fact that AI will actually play a solid role in strengthening our security posture. The report goes into a session at the recent Munich security conference where this topic was dealt in detail.Machine LearningThe biggest news in this edition in the Machine Learning space has been the release of Gemma, a family of lightweight, state-of-the art open models. These models that are based on the same technology as Gemini are available in two sizes: Gemma 2B and Gemma 7B. The key thing with these models is their open nature. Gemma models can run on a laptop, workstation, or on Google Cloud. You can build with and customize in Vertex AI and run it on GKE. Developers can use Gemma with their preferred tools/frameworks that includes Colab and Kaggle notebooks, JAX, PyTorch, Keras 3.0, and Hugging Face Transformers. Check out the article for more details.Imagine a scenario where you are watching a video of a famous chef preparing your favourite recipe. As the chef is mentioning the ingredients, they are automatically inserted into a shopping cart that you can then order right away. This is one such scenario that is possible when you combine the interesting world of Video Analytics with Generative AI capabilities. This article highlights this intersection and mentions interesting use cases with demos. It is sure to spark some ideas.DatabasesHave you been considering moving to Spanner PostgreSQL dialect with the announcement of its General Availability. If specifically you are looking to move from a PostgreSQL environment that you are managing, you should check out this detailed guide that shows not just the benefits of moving to it but a step by step guide to doing that, using the Spanner Migration Tool (SMT) that makes the task smooth.Data AnalyticsGoogle Cloud with its Data Analytics services can be difficult at times to combine together to form an overall solution. In an interesting Data Analytics solution that is worth studying, Deustche Bank built out a Trade Surveillance system. The system made good use of BigQuery and Dataproc services as shown in the diagram below. What’s interesting to note is how the problem was broken down into how the front office system would provide the relevant data on which the various surveillance scenarios were then run and were identified for investigation. Check out the article for more details.GamesIs there an area that Generative AI will not touch? Let’s talk about Games, where the article believes that it has a great potential in the following use cases: Game production, Adaptive gameplay and In-game advertising.The article also dives into Generative AI Framework for Games, which provides templates for running Gen AI for games on Google Cloud, as well as a framework for data ingest and storage to support these live models. The article covers a demo of the same.NetworkingA new internal range (IP net block for a VPC network) API aims to simplify management of private IP addresses. This API allows for allocate subnets from a free address range within your VPC, reserve and protect this range with a Google Cloud Network and reserve / protect on-premises, or cross-cloud ranges. Check out the article with various use cases on where this feature is useful.Developers and PractitionersThe last few months has seen Google release several updates to the Vertex AI models, starting with PaLM and then the Gemini models. You might have invoked these models from an API or SDK and even exercised some of the multi-modal features of Gemini. But what if you had a scenario where you are given a list of items (for e.g. countries) and you need to run the same query(prompt) on each of the item. As the article, rightly points out, you might end up with the context length issue if you send all the data in one prompt or you might end up sequentially invoking the model on each data. How about being able to do that in parallel and with an approach that can scale. The solution: Use Cloud Workflows as the article effectively demonstrates.The best way to get started with any technology is to take the services for a spin, integrate the functionality into your applications or build out prototypes of new applications. Customers have identified key Gen AI applications that have an immediate return for them. These include modernizing web sites, improving developer productivity, product cataloging and more. Check out this article that highlights each of these key areas and provides ready made repositories of these Gen AI applications that you can put to work together. If you are a Developer in this space and working with Google Cloud, this is a repository to study next.Learn Google CloudCross-Cloud Network is a solution that addresses the need from organizations to be able to build multi tier apps and network them regardless where ever they are. Key use cases for Cross-Cloud Network solution include distributed applications, content delivery, hybrid workforce solutions and more. Check out the article to learn more about this solution.Stay in TouchHave 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.Want to keep tabs on new Google Cloud product announcements? We have a handy page that you should bookmark → What’s new with Google Cloud.Google Cloud Platform Technology Nuggets — February 16–29, 2024 Edition was originally published in Google Cloud - Community on Medium, where people are continuing the conversation by highlighting and responding to this story.
[/content]

Author: Romin Irani
PublishedDate: 2024-03-01
Category: Blogs
NewsPaper: Romin Irani - Medium
Tags: tech-nuggets, gcp-weekly, google-cloud-platform