Kategorie: Google Cloud Platform

A cloud built for developers — 2021 year in review

2021 was a seminal year for software developers. Every company accelerated their digital and online efforts, while simultaneously moving to remote development. Innovation by driving developer productivity was top of mind for nearly every IT executive we spoke to. Many asked us about Alphabet’s long track record of innovation. From Google search to Waymo’s driverless cars,  is there a secret to developing the next big thing? The answer is simple: 10X thinking. Look for solutions that help customers drive 10X improvements, through a series of smaller increments that compound to a large impact over time. At Google Cloud, we follow a similar philosophy to help our customers become innovative technology companies. In recent times, we’ve worked closely with partners, customers, and developers on services that help unlock 10X improvements in developer productivity. Six years ago, we introduced a managed Kubernetes service, Google Kubernetes Engine (GKE). This year, we added GKE Autopilot, which revolutionized Kubernetes management by eliminating all node management operations. Likewise, our Cloud Run serverless platform was the first service of its kind, allowing developers to go beyond running small bits of code and run full applications in a serverless environment. From September 2020 to September 2021, Cloud Run deployments more than quadrupled. More recently, we co-founded the Open Source Security Foundation and began working on secure continuous Integration and delivery (CI/CD) services a year or so ahead of the cybersecurity threats that made it to headlines. Here are the top developer challenges that customers asked us to solve in 2021: Driving distributed developer productivitySecuring the software supply chainSimplifying running of cloud-native applications Read on for more insights. Driving distributed developer productivityA critical prerequisite for innovation is time. Investments in developer productivity free developers to work on the important things. Traditionally, developers have spent hours downloading and installing tools to their local environments, updating them with the latest versions, or dependencies. Cloud Shell Editor is a full remote development environment with a growing set of built in security capabilities. It comes with developer tools pre-installed, including MySql, Kubernetes, Docker, minikube, Skaffold, etc. Developers just needed a web browser and internet connection to be productive. Developers now have access to tutorials right from Cloud Shell Editor, and can try code samples directly in our documentation. Additionally, with support for buildpacks, developers can create container images directly from source code, without knowing anything about docker or containers. Securing the software supply chainSoftware supply chain vulnerabilities had far reaching consequences in 2021, with events such as SolarWinds, Mimecast/Microsoft Exchange, and Log4jaffecting businesses, daily life, and entire governments. President Biden even issued an executive order to strengthen software supply-chain security standards.Solving the software supply chain problem requires players across industries to work together. This is why we co-founded theOpen Source Security Foundation (Open SSF). We also proposed SLSA, an industry-wide framework for maintaining the integrity of software artifacts throughout the software supply chain. Open source, with its complex dependency trees, continues to remain a prime target for exploitation. In fact, an estimated 84% of commercial code bases have at least one open source vulnerability. Today, developers can use our tools such as Allstar GitHub App, open source security score cards and Open Source Insights to implement security best practices, determine a risk score for open source projects, and visualize a project’s deep dependencies. And several of these same  kinds of open-source innovations are available out of the box to Google Cloud customers. Here are a few examples: Detailed recommendations to help mitigate the Apache Log4j vulnerability. The Java scanning feature of Google Cloud On-Demand Scanning, which can be quite handy for developers to identify Linux-based container images that use an impacted version of Log4j. On-Demand Scanning can be used with no charge until December 31, 2021. Cloud Build, our serverless CI/CD service, offers SLSA Level 1 compliance by default. This verifiable build provenance lets you trace a binary to the source code to prevent tampering and prove that the code you’re running is the code you think you’re running. Cloud Build’s new build integrity feature improves on this by automatically generating digital signatures, which can be validated before deployment by Binary Authorization. Simplifying running cloud-native applicationsInnovation is rarely a straight road, there are many wrong turns along the way. Developers need a cost effective runtime, a way to run experiments and fail forward fast. That’s why GKE Autopilot takes GKE, the most mature Kubernetes service on the market and further simplifies Kubernetes operations by providing a managed control and data plane, an optimized configuration out-of-the-box, automated scalability, health checks and repairs, and pay-for-use pricing. “With GKE Autopilot, we can do more with our business. We can continue developing and upgrading our products, rather than focusing on fine-tuning infrastructure.”—Jun Sakata, Software Engineer, Site Reliability, Ubie Simpler still is no cluster all. Cloud Run provides developers the freedom to run services from code or container images with no cluster or VM to manage. At the same time, it provides a hypervisor grade secure sandbox environment and several built in DevOps capabilities such as, multi-versioned deployments, gradual rollouts and rollbacks, GitHub and Cloud Build integrations. This is ideal for web and mobile application development. In 2021, with additions like higher per-instance concurrency, new CPU allocation controls, and support for standard Docker images, the benefits of serverless can now be expanded to a wider range of workloads, including legacy ones. Additionally, with newer cost controls along with billing flexibility like committed use contracts and features like always-on CPU, it’s possible to run more steady-state pattern workloads cost effectively in a serverless environment.  Best of all, thanks to improvements like these, organizations using Cloud Run have reported reduction in developer recruiting costs by 40%. Cloud Run is also the first platform to provide developers the option to optimize their carbon footprint.  With the news self-service Region Picker you can choose the data center region with the lowest gross carbon cost on which to run your Cloud Run workloads. Further, with just one click, Google Cloud Carbon Footprintgives you access to the energy-related emissions data for external carbon disclosures. “With Cloud Run, we only need half the people to manage our systems as compared to before” Google Cloud Platform Architect, Cosmetics “Cloud Run is one of the easiest services on Google Cloud Platform you can deploy to. It’s just super simple.” CTO,Healthcare SaaSIf you want to give Cloud Run and associated Cloud Functions a try, check out the Easy as Pie Serverless Hackathon, which offers  over $20,000 USD in cash prizes.2022: More to come  2021 brought simplification and greater attention to developer productivity. It is essential that developers continue to operate at even higher levels of the stack, without worrying about infrastructure, security, compliance and integrations. This is the Northstar for 2022. In 2022, look for Google Cloud to co-innovate with our ISV partners, developers, and SecOps team to bring you the 10X innovation you need from the cloud that is built for developers.
Quelle: Google Cloud Platform

Cloud CISO Perspectives: December 2021

This is our last Cloud CISO Perspectives of 2021. It’s been an eventful year for the cybersecurity industry, both good and bad, and I welcome the opportunities and challenges we will continue to address together in 2022. In this final post, I’ll share the latest updates from the Google Cybersecurity Action Team, new reports from Google’s security research teams and more information on Google Cloud’s Log4j impact and assessment.Update on Log4j vulnerabilityGoogle Cloud continues to actively follow the evolving security vulnerabilities in the open-source Apache “Log4j” utility and we are providing regular updates to our security advisory page. Responding to these vulnerabilities can be especially stressful, even more so when reaching the end of the year. We encourage everyone using vulnerable versions of Log4j, in any environment, to upgrade as soon as possible and according to guidance published by Apache, found here. As the entire industry works through its response to Log4j, the Google Cybersecurity Action Team also continues to publish and update recommended actions for mitigating exposure to the Log4j vulnerabilities. The state of open source software securityWhat recent events have taught us and will continue to teach us into 2022 is that we owe our thanks to the volunteers and maintainers of open source software. More than ever, we need continued industry investment and commitment to support them.For years, Google has been focused on addressing this challenge. Our open source security team helped found the Open Source Security Foundation (OpenSSF). Over the past year, we have doubled down on our investments in open source software security; from tools to frameworks to funding maintainers of open source software projects to focus on security. This past August, we committed $10 billion to advancing cybersecurity for organizations and governments globally where a major part of that commitment is focused on securing the open source software ecosystem, including $100 million in investments to third-party organizations like Linux Foundation and OpenSSF. One of the primary challenges facing defenders at this very moment is simply getting a handle on where Log4j dependencies exist within their organization’s codebases. Our Supply-chain Levels for Software Assurance (SLSA) project, which we open sourced in partnership with the OpenSSF, is an end-to-end framework to manage supply chain integrity and security, and its implementation would greatly aid organizations in this kind of situation. Last week, Google’s Open Source Insights team published an analysis on the impact of the Apache Log4j vulnerability where they pulled together a list of 500 affected packages with some of the highest transitive usage and encouraged maintainers or users helping with the patching effort to maximize impact and unblock more of the community. Improvements such as these could qualify for financial rewards from the Secure Open Source Rewards program. You can explore your package dependencies and their vulnerabilities by using Open Source Insights. We all can do our part to support this critical function of our software ecosystem, and I look forward to seeing how organizations, governments and individuals work together to make improvements in the coming year. Google Cybersecurity Action Team Highlights Below I’ll recap the latest updates, new services and resources across our Google Cybersecurity Action Team, Google Cloud Security product teams and Google security research efforts since our last post. SecurityQ4 Cloud Security Talks Recap: We hosted our final Google Cloud Security Talks event of 2021 where our security teams focused on zero trust and covered everything from Google’s history with BeyondCorp to our strategic thinking when it comes to applying zero trust principles to production environments. We also shared product updates across the portfolio and talked about how zero trust fits into our invisible security vision. Check out the recap in this blog post and watch the sessions virtually on-demand.Autonomic Security Operations: Our Autonomic Security Operations solution continues to resonate with organizations and security professionals widely as teams look for more ways to modernize their security operations. Dr. Anton Chuvakin and Iman Ghanizada from the Google Cybersecurity Action Team recently published a whitepaper on how organizations can work towards a 10x transformation of their SOC. Their first blog post in a series of many looks at what security teams can learn from Site Reliability Engineering (SRE) principles and philosophies to begin their journey towards modernizing the SOC. ComplianceSoftware-Defined Community Cloud: Our Google Cloud compliance team outlined a new concept for how the industry can address challenges within legacy community cloud implementations. Our Assured Workloads product implements a novel approach to help customers meet compliance and sovereignty requirements through a software-defined community cloud. A software-defined community cloud is designed to deliver the benefits of a community cloud in a more modern architecture. Google Cloud’s approach provides security and compliance assurances without the strict physical infrastructure constraints of legacy approaches.Continuous Compliance: Following the Google Cybersecurity Action Team’s launch of the Risk and Compliance as Code solution, our customer engineering teams shared some timelycase studies on how Google Cloud customers are reaching continuous compliance, encompassing real-time attestation and notification. The key learning: the more familiar control owners become with our GCP capabilities, the more confident they feel to automate their controls.Shared FateSecured Data Warehouse blueprint: Google Cloud customers can jump start the migration and analysis of sensitive business data by using the new Google Cloud Secured Data Warehouse blueprint. This opinionated guidance consists of both documentation and deployable Terraform assets.  It is built around BigQuery and incorporates Cloud DLP, Cloud Storage, PubSub, Dataflow, Data Catalog, and CMEK to implement security best practices across data ingestion, storage, processing, classification, encryption, logging, monitoring and governance. Security Foundations Blueprint v2.5: And we’re excited to announce the next version of our Security Foundations Blueprint. New content provides further control for data residency and also supports Assured Workloads for enhanced native platform guardrails.  We review the guide and corresponding blueprints regularly as we continue to update best practices to include new product capabilities. Controls and ProductsNetwork-based Cloud threat detection with Cloud IDS: We announced the general availability of our Cloud IDS solution that helps enterprises detect network-based threats and helps organizations meet compliance standards that call for the use of an intrusion detection system. With the general availability, Cloud IDS now has the following enhancements: service availability in all regions, detection signatures automatically updated daily and new compliance support for customers’ HIPAA compliance requirements and ISO27001 certification.New zero trust features in BeyondCorp Enterprise: The BCE team released the Policy Troubleshooter feature in general availability. The tool provides support for administrators to triage blocked access events and easily unblock users within an organization, which is an essential tool for admins as employees continue to work remotely or in hybrid and need ways to access corporate resources and information securely. Keyless Authentication from GitHub Actions: Following GitHub’s introduction of OIDC tokens into GitHub Actions Workflows, you can now authenticate from GitHub Actions to Google Cloud using Workload Identity Federation, removing the need to export a long-lived JSON service account key.  New functionality like this is a part of Google Cloud’s ongoing efforts to make security invisible and our platform secure-by-default. Learn more in the blog post. Threat Intelligence Combating cyber crime at scale: In December, Google took action to disrupt Glupteba, a sophisticated botnet targeting Windows machines. This was also the first lawsuit against a blockchain enabled botnet, where the attackers protected itself using blockchain technology. Google’s Threat Analysis Group took steps to detect and track Glupteba’s malicious activity over time and we launched litigation which we believe will set a precedent and help deter future activity. The details in TAG’s analysis and our litigation demonstrate that crime on the internet is sophisticated, and at Google, we feel a responsibility as part of this ecosystem to play a part in disrupting this activity to help everyone on the Internet be safer.iMessage zero-click exploit: In a recent blog post, Google’s Project Zero researchers show for the first time how an in-the-wild zero-click iMessage exploit works and how it is used by NSO. Must-listen podcasts Earlier this month, the Google Cloud Security podcast hit a major milestone: 46 episodes in its first year! Check out this post to see the top themes from our podcast throughout 2021, including episodes on zero trust security, cloud threat detection, how to make cloud migrations more secure and data security in the cloud. This wraps up the year for Cloud CISO Perspectives in 2021! We’ll be back in 2022 with continued updates from our Google Cybersecurity Action Team and more. If you’d like to have this Cloud CISO Perspectives post delivered every month to your inbox, click here to sign-up.Related ArticleCloud CISO Perspectives: November 2021Google Cloud CISO Phil Venables shares his thoughts on the latest security updates from the Google Cybersecurity Action Team.Read Article
Quelle: Google Cloud Platform

Google Cloud’s top AI blog posts from 2021

Artificial intelligence (AI) remained in the spotlight over the last year, as the gap continued to grow between organizations that merely possess data and those that can use data to leverage the power of AI to generate actionable insights or improve customer experiences. At Google Cloud, helping you turn AI investments into real-world results is one of our foremost goals—and we kept our foot on the gas in 2021, launching a variety of new solutions, research, and tutorials. But don’t fret if you might have missed anything along the way. Whether you’re a seasoned data scientist or someone looking to solve problems with AI for the first time, Google Cloud offers platforms, tools, and best practices for all levels of expertise—and to close out the year, we’ve collected some of our top 2021 AI blog posts to help you kick off 2022 on the right foot.  Vertex AI: one platform for all your ML toolsUnveiled in May, Vertex AI was one of our most significant AI announcements of the year. A managed machine learning (ML) platform, Vertex AI supports your data teams more quickly building, deploying and maintaining ML models. Compared to competing platforms, it requires almost 80% fewer lines of code to train a model, helping your organization to implement Machine Learning Operations(MLOps) across all levels of expertise. Whether you’re newly adopting Vertex AI in 2022 or have been using it for months, here are a variety of articles to help you take full advantage of its powerful feature: What is Vertex AI? Developer advocates share moreAI Simplified: Managing ML data sets with Vertex AIUse Vertex AI Pipelines to build an AutoML classification end-to-end workflowBuild a reinforcement learning recommendation application using Vertex AIVertex AI Matching Engine: Blazing fast and massively scalable nearest neighbor searchPyTorch on Google Cloud: How To train and tune PyTorch models on Vertex AIAnnouncing Vertex AI Pipelines general availabilityVertex AI NAS: higher accuracy and lower latency for complex ML modelsCoca-Cola Bottlers Japan collects insights from 700,000 vending machines with Vertex AIGoogle demonstrates leading performance in latest MLPerf BenchmarksCCAI: reimagining customer experiences with the power of AIGoogle Cloud’s Contact Center AI (CCAI) platforms make ML-powered language models more accessible and impactful by helping even companies with limited AI expertise to uncover insights in customer and partner interactions, and deploy virtual agents that can chat naturally with customers.CCAI is built specifically to help call centers deliver excellent customer service on demand, even when human agents are unavailable, and we added powerful new features throughout the year, including Speaker ID, whichlets customers authenticate themselves with just a few spoken words; Agent Assist, which provides human agents with continuous insight into customer intent during calls and chats; and CCAI Insights, which uses AI to mine raw contact center interactions for insights, regardless of whether the data originated with a virtual or human agent. To learn more about how customers are leveraging CCAI, don’t miss our post about HSBC’s use of Dialogflow, a part of CCAI, to ease the call burden on its policy experts. DocAI: unlocking the value in unstructured data Our Document AI (DocAI) platform eliminates the guesswork and manual labor involved with document processing, helping your teams to better understand the data captured in documents and to streamline workflows. So your business can move even faster from implementation to value, we’ve introduced use case-specific additions to the platform, including Lending DocAI, Procurement DocAI, and Contract DocAI. To dive into how our customers are putting DocAI solutions to work, be sure to check out these articles: How Mr. Cooper is using AI to increase speed and accuracy for mortgage processingGoing global: Workday uses Google Cloud AI to accelerate document processingIndustry solutions: smarter decision making, from retail to manufacturing Delivering the right information to customers in the right context is crucial, so we’re pleased to see the great results that retail customers like IKEA and Bazaarvoice have enjoyed with our Recommendations AI solutions, achieving 30% and 60% increases in click-through rates, respectively. We published several blogs in 2021 to help you do more with this solution, including: Recommendations AI data ingestionRecommendations AI modelingServing predictions & evaluating Recommendations AIHow to get better retail recommendations with Recommendations AIIn addition to our use case-oriented DocAI solutions and our Recommendations AI work with retailers, we’ve also been digging into other specific industries and challenges, highlighted by the following:Visual Inspection AI: a purpose-built solution for faster, more accurate quality controlNew research reveals what’s needed for AI acceleration in manufacturingTranslation: customer connections without language barriersTranslation is one of the fastest growing AI use cases, and we rolled out a wide range of feature updates to help you connect with customers, highlighted by the ability to translate business documents across more than 100 languages. If faster translation workflows are among your 2022 resolutions, don’t miss our post about best practices for translating websites with Translation API or our article about how the city of San Jose uses AI translation to ensure critical services reach the community. We’re just getting started: training and recognition At Google Cloud, we see AI continuing to impact business and continuing to become easier to implement and leverage—so the preceding 2021 are just the tip of the iceberg. We’re looking forward to helping your organization do even more with AI in 2022, but in the meantime, here is a collection of blogs highlighting some of our additional research, recommendations, and plaudits, as well as training resources to help you do more with AI, faster: Forrester names Google Cloud a leader in AI InfrastructureGartner names Google a leader in 2021 Magic Quadrant for Cloud AI Developer Services reportGoogle Cloud AI leaders share tips for getting started with AIWhy you need to explain machine learning modelsFree AI and machine learning training for fraud detection, chatbots, and moreGrow your ML skills with free offer from Coursera7 tips for trouble-free ML model trainingYour guide to all things AI & ML at Google Cloud NextRelated ArticleGoogle Cloud expands CCAI and DocAI solutions to accelerate time to valueGoogle Cloud deepens customer understanding with Contact Center AI Insights and transforms contract management with Contract DocAIRead Article
Quelle: Google Cloud Platform

The top three insights we learned from data analytics customers in 2021

As you know, we’re obsessed about learning from customers. That’s why, we made it a point to sit down with a new Google Cloud customer every week this year to share with you their journey both to our unified data cloud platform and what they’ve learned from their experience along the way. What began as a simple exercise of routinely keeping up with our customers quickly evolved into a complete video series that I have the pleasure of hosting every Tuesday. This year you learned from companies of all sizes, all industries and across the globe.In this series, we learn that succeeding with data requires tough business decisions, a lot of creativity and a specific vision. Perhaps for the first time, this series, aptly titled “Data Journeys,” spotlights our community voices from the data space in a genuine, grassroots, and unfiltered way.We’ve learned so much from each customer guest and their company’s unique challenges, trials, and triumphs – all things that make up a great journey or adventure. So, as we put the wrappings on another eventful year, we’re thrilled to tie a bow around our 2021 customer “Data Journeys” and share the top 3 learnings from their journeys with you.Lesson 1: Customers migrating to the cloud aren’t simply looking for a provider that can successfully execute a lift-and-shift motion. Instead, they look to take advantage of their cloud migration as an opportunity to rethink, relearn, and rebuild. When they were reaching the end of their contract with a different data warehousing provider with a fixed compute model, Keybank, a regional financial institution with over 1,000 branches managing over $145 billion in assets, decided to migrate to the cloud. I was fortunate to speak with Mike Onders, EVP Chief Data Officer, Divisional CIO, and Head of Enterprise Architecture for Keybank who led the migration charge. According to Onders, he decided Google Cloud Platform was the right fit for the dynamic, security-sensitive nature of the banking company’s business because Google’s data suite provided an elastic, fast, and consistent environment for data. When selecting Google Cloud however, Keybank was clear they did not want to simply lift-and-shift their Teradata Warehouse, analytics users, and Hadoop Data Lake to Google and wipe their hands to call it a day. Instead, they saw the migration as an opportunity to reinvent old processes and traditional ways of doing banking. “We really spun up Python and Spark clusters to do some new fraud modeling, especially around smart holds and check deposits and how to use fraud algorithms to determine whether we should hold this money or release it to you when you deposit that check,” Onders said.We really spun up Python and Spark clusters to do some new fraud modeling, especially around smart holds and check deposits and how to use fraud algorithms to determine whether we should hold this money or release it to you when you deposit that check Mike Onders , EVP Chief Data Officer, Divisional CIO, and Head of Enterprise Architecture at KeybankIn addition to smart check hold modelings, Keybank has re-engineered how they conduct attrition modeling, next best offers/product recommendations, and credit risk predictions using Google Cloud’s intelligent data cloud platform. Now, they are eyeing even more new innovation, particularly around how to use data to drive customer satisfaction and differentiate their services from their competitors. .For more insights, watch Keybank’s episode on Data Journeys. Lesson 2: Customers are passionate about applying Google’s technology, from predictive analytics to AI and automation, to ‘data for good’ initiatives.TELUS is one of Canada’s largest telecommunications companies with over 9.5 million subscribers. The powerhouse manages geolocated data generated by talk and text signals, networks, and cellular towers across the country. For some perspective on the sheer amount of data they handle, TELUS analyzed over 1.2 petabytes of data last year and are expecting that number to grow more each year. In March of 2020 it became clear that COVID-19 was more than an outbreak and growing to be a global pandemic, and TELUS faced a choice to carry on business as usual or to serve their community at a larger scale. With Google Cloud by their side to support them, the choice was a no brainer.TELUS realized they could leverage their data residing within Google Cloud’s data cloud in a privacy-preserving manner to launch a data for good initiative and support all Canadians during a time of need. Within a matter of just 3 weeks using tools like BigQuery and Data Studio, TELUS launched a new platform that empowered the Canadian government to make better strategic decisions about how to combat COVID within its borders – all while keeping individual subscribers’ data private.For their efforts, TELUS earned a HPE-IAPP Privacy Innovation Award and helped the government contain and minimize the spread of the virus.After speaking with Michael Ames, Senior Director of Healthcare and Life Science at SADA, I realized their company was similarly focused on using data analytics technology for community good. SADA is a Google Cloud Premier Partner that offers consultation and implementation services. Ames’ day-to-day involves helping hospitals and healthcare providers migrate their data to the cloud. Hospitals and healthcare systems around the United States undoubtedly handle massive amounts of data from symptom records to patient history to insurance and payment records, and more. SADA realized that increasing the rate that the right treatment would be matched with the right person during their first visit would help hospitals run their business better, increase patient satisfaction and health, save both parties time and money, and therefore maximize profits that hospitals could reinvest in care for their community. But, after conducting a study with the journal Nature taking the top ten biggest drugs by revenue and analyzing how they affected patients to determine whether they achieved their intended purpose, Ames knew matching the right drug with the right patient was a major issue. Once SADA used data to create predictive models within Google’s intelligent data cloud stack, they were able to help doctors prescribe medications that better matched individual patient needs and health profiles, providing far more personalized medicine. Looking forward, SADA is excited to use a data-driven approach to improve other aspects of health care. “Take that idea of prescribing drugs and expand that to physical therapy treatments, behavioral therapy treatments, decisions we make about where and how people should live, whether we should build a park in a certain neighborhood, how the air quality is affecting health, and we start to get a sense of the need to bring data together in a way that will be… much more powerful in treating the health of people” Ames said.Take that idea of prescribing drugs and expand that to physical therapy treatments, behavioral therapy treatments, decisions we make about where and how people should live, whether we should build a park in a certain neighborhood, how the air quality is affecting health, and we start to get a sense of the need to bring data together in a way that will be… much more powerful in treating the health of people Michael Ames, Senior Director of Healthcare and Life Science at SADAFor more insights, watch TELUS’ episode and SADA’s episode on Data Journeys. Lesson 3: Customers choose Google Cloud to address today’s needs, but also because they are gearing up with technology that can solve their unknown problems of tomorrow. Delivery Hero is the largest food delivery network outside of China. The food delivery company operates in over 50 countries and processed over 663 million orders in just the first quarter of 2021.I was shocked to learn that millions of orders equal 174 data sets, 5,000 tables, and upwards of 7 million queries a month. Hence, it was my pleasure to speak with Matteo Fava, Delivery Hero’s Senior Director, Global Data Products and Analytics, and understand how they manage their immense data volume.A significant part of Delivery Hero’s data strategy involves making sure the right teams have access to the right data on the backend, and that their customers have a safe, efficient, and enjoyable experience of receiving their food on the frontend. In fact, that was the main reason why they partnered with Google Cloud in the first place: to ensure business would run as usual with no hiccoughs.  As Fava’s team gained deeper understanding into user behavior and delivery routes through Google Analytics and BigQuery, Delivery Hero discovered a puzzle they hadn’t previously anticipated of how to best deliver food in a consistent way despite significant cross-cultural differences between countries or even cities or neighborhoods within the same country. One such puzzle was how to optimize “horizontal” deliveries such as delivering to standard homes versus “vertical” deliveries such as delivering to apartment buildings while often keeping a 15 minute delivery promise to their customers. Delivery Hero ultimately used real-time insights into deliveries and predictive modeling to ensure their estimated time of food delivery took into account longer elevator wait time in populated cities like Hong Kong and the time it takes to enter and navigate condominium complexes in places like Dubai. Delivery Hero knew it was one thing to have an efficient new platform for managing data, but it’s another to trust that this platform can solve unexpected challenges that may not be apparent today. For more insights, watch Delivery Hero’s episode on Data Journeys. Wrapping it upWe can’t wait to explore new customer data journeys next year and continue to share insights with the  community. We hope you’ll follow the Data Journey series  by subscribing to the  playlist, and if you have a great suggestion for a guest (or want to be a guest yourself), please let us know!
Quelle: Google Cloud Platform

Google Cloud enables the National Cancer Institute's Institute for Systems Biology-Cancer Gateway in the Cloud to support breast cancer research with fast and secure data sharing

Research organizations today recognize the challenge of sifting through siloed data sets, and analyzing and sharing this data with the global research community—all while staying secure and compliant within a range of national and international standards. It is precisely these constraints that led the U.S. National Cancer Institute (NCI) to create Cloud Resources, which are components of the NCI Cancer Research Data Commons that allow scientists to analyze cancer datasets in a cloud environment (vs. having to download data and use custom hardware). Included in these resources is the Institute for Systems Biology-Cancer Gateway in the Cloud (ISB-CGC). ISB-CGC relies on Google Cloud to securely host terabytes of genomic and proteomic data, and provide flexible and scalable analytics tools that can be integrated into research models. Complex computations that traditionally required days to complete are now executed in just minutes or hours. And ISB-CGC can now deliver open data, compute, and analytics resources to the global research community.Enabling faster time-to-discoverySpeed and scale can make all the difference when it comes to potentially life-saving research. Take breast cancer for example. It’s the world’s most prevalent cancer and according to the World Health Organization, more than two million women were diagnosed with it in 2020 alone. With such a large number of impacted women, each with unique biological features and personal paths through the disease, breast cancer research is particularly data intensive. And processing this on-premises is too slow, expensive, and burdensome to patients.By working with Google, NCI’s ISB has not only made data more useful to cancer researchers around the world, but also has fundamentally changed how cancer investigators conduct research. BigQuery, Google Cloud’s highly scalable mulitcloud data warehouse, underpins the cloud-based platform that connects researchers to a wide collection of cancer datasets, as well as the analytical and computational infrastructure to analyze that data quickly. “We are spreading the message of the cost-effectiveness of the cloud,” said Dr. Kawther Abdilleh, lead bioinformatics scientist at General Dynamics Information Technology, a partner of ISB. “With Google Cloud’s BigQuery, we’ve successfully demonstrated that researchers can inexpensively analyze large amounts of data, and do so faster than ever before.”Integrating diverse tools and datasetsTraditionally, researchers have downloaded source data and performed analysis locally on their personal machines using programming languages like R and Python. As the volume and complexity of cancer data has grown, this approach has become unsustainable. Through the use of Google Cloud services, like Notebooks and BigQuery application programming interfaces (APIs), researchers can now use their desired methods to analyze data on the ISB-CGC platform, directly in the cloud—without the need to download data. For example, in their September 2020 paper on data integration and analysis in the cloud, Dr. Abdilleh and Dr. Boris Aguilar, senior research scientists at ISB, demonstrated how cloud-based data analysis can be used to identify novel biological associations between clinical and molecular features of breast cancer. “Google’s AI platform, for example, allows us to easily create notebooks to use R or Python in combination with BigQuery or machine learning to perform large-scale statistical analysis of genomic data, all in the cloud,” Aguilar wrote. “This type of analysis is particularly effective when the data is large and heterogenous, which is the case for cancer-related data.” Drs. Abdilleh and Aguilar developed a set of BigQuery user-defined functions (UDFs) to perform statistical tests and gain a more holistic picture of breast cancer. Performing these statistical functions directly on the massive data stored in BigQuery vs. in an on-premises computer program later in the analysis workflow saved a significant amount of time. In fact, by using UDFs with BigQuery, analysis that typically required supercomputers and days of computation was complete in minutes. Drs. Abdilleh and Aguilar have now made their UDFs available for use by the broader research community via BigQuery, opening doors for fellow breast cancer researchers to build on this progress and make strides in their life-saving work.  Global access to critical cancer dataWith so many lives and families impacted by cancer—and researchers worldwide diligently seeking answers—the need to accelerate and improve the means by which cancer research is conducted is critical. ISB-CGC’s success using Google Cloud as the foundation of its infrastructure and data cloud strategy has opened the door for the cancer research community to gain real-time, secure access to data that plays a significant role in the early detection of cancer. Read the case study for more detail on how Google Cloud is supporting breast cancer research.
Quelle: Google Cloud Platform