Google Cloud Platform Podcast

Matthew Lawson of IKEA Retail (Ingka Group) joins Mark Mirchandani and Priyanka Vergadia today, telling us all about IKEA Retail (Ingka Group)’s move to the cloud. Engineering Manager Matt and his team primarily focus on the early stages of development at IKEA Retail (Ingka Group), helping the company with research and planning as well as development. Lately, they have been focused on incrementally moving IKEA Retail (Ingka Group)’s digital presence to the cloud. Matt explains the digital shift process for IKEA Retail (Ingka Group) and why they chose to modernize and move pieces to the cloud over time. By illustrating through examples, he details projects the team worked on during this digital transformation. Matt also talks about the changes to the IKEA Retail (Ingka Group) Digital DNA, emphasizing progress made in their digital culture to allow for the drastic change from on-prem to the cloud. Using managed services like Google Cloud Run, IKEA Retail (Ingka Group) has been able to adapt and grow in the clou

Max Saltonstall is back in the co-host seat, joining Mark for a fun chat about Active Assist. Chris Law is our guest this week, and he starts with the story of his arrival at Google and his path to the Active Assist team. Active Assist is Google’s way of helping clients learn about and take advantage of all the cloud features available. Chris describes the main ways Active Assist helps customers in the real world, from troubleshooting tools that identify problems to analysis software that helps clients determine how changes will effect the project. Active Assist also provides recommendations for cost savings, better security, and performance boosters to help clients proactively build better projects. As companies scale, these features become even more important, Chris tells us, citing examples from real users. Later, we talk about how machine learning is employed to create these recommendations. We talk transparency and learn how Chris and his team keep open communication with clients as they design and impro

This week on the podcast, we’re celebrating GKE’s fifth birthday! Mark Mirchandani is joined by special guest host Carter Morgan to talk all things Kubernetes and GKE with fellow Googlers and GKE experts Alex and Drew. Drew starts the show with a thorough explanation of Kubernetes, telling our hosts that its a great way to manage containers as you scale. Because it is an open source offering, Kubernetes has grown and adapted quickly. Alex elaborates, pointing out that Kubernetes has helped redefine how people create cloud native applications. A year after Kubernetes was born, Google introduced Google Kubernetes Engine to help simplify things for developers while optimizing scalability and efficiency. Our guests talk about the progression of GKE over its short life, what’s new with the latest version, and why reliability and scalability have become the focus for year six. Later, we hear examples of companies taking advantage of everything GKE has to offer and how the symbiotic relationship between Google and i

Mark and Brian Dorsey are together again this week as we learn all about Google’s Bare Metal Solution with our guests James Harding and Gurmeet “GG” Goindi. To start the show, GG introduces us to Bare Metal Solution, explaining that it allows client projects built on specialized, often outdated software to take advantage of the benefits of a cloud environment. Using Bare Metal Solution, clients can choose to migrate all or part of their projects for a fully customized experience. We learn how Bare Metal Solution is able to support a partial or full native solution for clients and go through the steps to getting a project from completely on-prem to the cloud where latency is decreased, security is increased, and other cloud benefits can be leveraged. GG gives examples of situations where Bare Metal is a great option for clients, for instance an established company with an early 90s database that recently branched out into apps built in cloud native software. James outlines the benefits of Bare Metal Solution o

This week on the podcast, Mark and Max Saltonstall talk with Simen Svale Skogsrud and Knut Melvær of Sanity.io. Sanity.io started as a consulting company but organically morphed into a software company when they realized their content management solutions worked across many industries. By providing a managed system that includes search indexing and data hosting, Sanity.io allows customers to analyze and deliver content all over the world with ease. They also offer an open source kit that facilitates complete customization of the program to each client’s particular needs. Simen explains headless CMS as compared to the conventional systems and how it benefits Sanity.io clients. Data is separate and much more flexible, allowing it to be used in any way on any platform. Knut tells us about the developer experience using Sanity, describing the dashboard of useful APIs and other features that make using the program a breeze. We talk about how real clients have influenced and built on the product and why customer se

Priyanka is back this week, joining Mark as we talk big data with our guests Prabhu Palanisamy and Jarrett Garcia. Iron Mountain, a data management company, securely stores hard-copy and online data for enterprise customers. SpringML aides enterprises in the use of data analytics and machine learning to transform their projects. With the help of SpringML and Google Cloud, Iron Mountain migrated much of their data to datalakes and cloud storage for easier access and manipulation of the data. Jarrett and Prabhu talk about the process of migrating so much data, including the main goals of their partnership. First, they established systems that could be repeated and help Iron Mountain understand when data is moved, destroyed, migrated, and more. We discuss the next steps taken, learning how Iron Mountain moved so much data to the cloud. Using Google products like DataFlow, BigQuery, and Cloud Composer, SpringML was able to take the data from Iron Mountain’s landing zone and transform it. Prior planning meant that

On the podcast this week, Mark Mirchandani and Brian Dorsey talk with fellow Googlers John Laham and Stewart Reichling about Traffic Director, a managed control plane for service mesh. Traffic Director solves many common networking problems developers face when breaking apart monoliths into multiple, manageable microservices. We start the conversation with some helpful definitions of terms like data plane (the plane that data passes through when one service calls on another) and service mesh (the art of helping these microservices speak with each other) and how Traffic Director and the Envoy Proxy use these concepts to streamline distributed services. Envoy Proxy can handle all sorts of networking solutions, from policy enforcement to routing, without adding hundreds of lines of code to each project piece. The proxy can receive a request, process it, and pass it on to the next correct piece, speeding up your distributed system processes. But Envoy can do more than the regular proxy. With its xDS APIs, service

Mark Mirchandani is joined again by Priyanka Vergadia this week for an ML-filled interview with Radu Miclaus of Lucidworks. Lucidworks, a company specializing in information retrieval, strives to make data searching easier for developers and users. Building off Solr, Lucidworks created Fusion, an environment more conducive to easy AI-enhanced query capabilities, better scalability, and more. With Fusion, developers can take advantage of the highly advanced relevance tuning tools such as query rewrites, which analyze user behavior and automatically rewrite queries based on that information. On the tech side, Fusion was built with a combination of Java, Kubernetes to increase scalability, Solr management tools, and logging and reporting tools. The engineers at Lucidworks have created Fusion-specific system-enhancing pieces as well, including a machine learning service that allows data scientists to train their models elsewhere and plug them in for a completely customized experience. The team also created Smart

Tyler McMullen of Fastly is with us today, telling our hosts Mark Mirchandani and Brian Dorsey all about the company, CDNs, and more. Fastly is an edge cloud platform, focusing on ways to improve the more customer-focused side of the cloud with things like latency reduction, efficient scaling, and more. Content Delivery Networks can be a part of this, due to their proximity to customers and better caching. Edge cloud takes pieces of normal cloud setups and moves them to the edge of the cloud, closer to the customer, to achieve better speed. Tyler explains what pieces make sense to move out to the edge and what he sees as the future of edge cloud platforms. Later in the show, Tyler tells us how to analyze projects and make decisions on the use of edge cloud, CDNs, and microservices. He explains the technical process of using an edge cloud platform too, giving examples of situations that might benefit from a more edge cloud approach. WebAssembly, technology originally created for web browsers, actually plays a

This week is all about business-to-business marketplace software with Pandium as Mark Mirchandani and Max Saltonstall talk with our guests Cristina Flaschen and Kelly Sarabyn. The Pandium platform helps companies build and support in-app marketplaces with a focus on software integration and flexibility. Kelly and Cristina start by explaining how Pandium deals with scalability for clients with multiple users and partners. Cristina elaborates on Pandium’s role in facilitating integrations, helping customers build customized, flexible solutions. We discuss how APIs are handled and the way Pandium takes care of authentication, security, and other standard pieces. We continue with a thorough discussion of ‘yes code’, ‘low code’, and ‘no code’ approaches, and the benefits and drawbacks of each system. With a combination approach of some ‘no code’ tools and other ‘yes code’ pieces, Pandium allows better customization in any code language, while keeping some functions easy for non-engineers. Cristina Flaschen Cristin

Mark Mirchandani and Priyanka Vergadia are here this week to talk about some cool things that are going on. Then, Mark brings us an interesting interview all about documentation in development with Technical Writing Manager Riona Macnamara. Riona specializes in technical documentation for DevOps at Google. Having written technical documentation for both external and internal audiences, Riona starts the show comparing the two, explaining how the process can be very different. In external writing, she stresses the importance of engineers and technical writers working together. She details the challenges of documentation in code development culture and offers some solutions. Laying out goals can create better quality documents, while providing a simple documentation process for engineers can help bolster a culture of documenting. She talks about open source projects, stressing that documentation is possibly more necessary because of the diverse contributor base. We wrap up the interview learning how to get start

This week, Mark Mirchandani and Priyanka Vergadia learn all about Cloud Audit Logging with fellow Googlers Philip O’Toole and Oscar Guerrero. Our guests explain the importance of Cloud Audit Logs to keep track of your GCP resources so you know who, what, where, and when things were done. Our guests explain the types of logs GCP offers and why each is important for security. The interview continues with a discussion the various other benefits of audit logging, including proof of compliance and measuring of risk. Because audit logs have the ability to create more data than some businesses can use, Philip and Oscar help our listeners understand how to choose the correct logging services for their needs, and we learn how Cloud Logging can help users digest their data. Philip describes how audit logs and event driven systems can benefit businesses, explaining how event driven systems can be built and pushed with GCP. Oscar continues the conversation with audit logging in G Suite. The Cloud Logging team is continui

Mark Mirchandani and Priyanka Vergadia host this week’s episode of the podcast, with a thorough discussion of Solutions Engineering at Google. Our guests, Grace Mollison and Ann Wallace, explain that the Solutions Engineering team is there to help customers choose appropriate products for things like security, analytics, data management, and more. The products are laid out in guides and blueprints so the client can easily understand why products are chosen and how to use them. Grace and Ann talk later in the podcast about the Solutions Engineering blueprints that Solutions Engineering Architects have begun creating. They describe how the idea came about, how they’re built, as well as the types of blueprints that are available and how to use them. The team is still working to create more blueprints and make them even easier to use. Grace Mollison Based in London, UK, Grace Mollison leads the Cloud Solutions Architect team in EMEA, where she helps customers to understand how to architect and deploy applications

Mark Mirchandani is back this week as he and co-host Brian Dorsey learn all about voice coding with some great guests! Emily Shea, senior software engineer at Fastly and user of Talon Voice, and Ryan Hileman, developer of Talon Voice, tell us about Repetitive Strain Injury and how it led to the design and use of Talon Voice. Talon allows not only straight voice input but includes eye tracking, noise recognition, and user customization to perform complex actions like moving windows, selecting text, and user-specific workflow customizations without touching any hardware. Emily describes her experiences with Talon, including the process to get started and how she looked past voice recognition stereotypes to find how useful the product could be. She demos Talon for us, explaining how the alphabet system works. Ryan types a sentence using his voice then explains the process of developing the alphabet and other parts of Talon. Later, Ryan and Emily tell us how they write code using Talon and the logistics of using

We’re discussing security operations on the podcast this week with your hosts Priyanka Vergadia and Mark Mirchandani. They’re joined by Elliott Abraham and Jason Bisson who start the interview explaining that they created the CLAM framework to help customers use Google Cloud security features to their fullest potential to create safe projects and relaxed clients. The CLAM (Cloud Logging Alerting and Monitoring) framework came about specifically to help customers transition products to, and run products securely in, the cloud. Using the Mitre GCP Matrix, the security team addressed each element with GCP product solutions, from initial access to persistence and beyond. CLAM is GCP specific, taking into account the default security measures GCP already provides and supplementing these measures with appropriate procedures for each client. Once the framework is in place and things are secure, clients can build on that with operational controls, such as SRE best practices. Elliott explains the shared security model

On this episode of the podcast, our old pal Mark Mirchandani is joined by special guest host Max Saltonstall to talk trust and security with fellow Googler Rob Sadowski. BeyondCorp is Google’s answer to allowing employees to use company networks on any device while outside the building in a way that is both secure and efficient. Users are authenticated per session and per device to give access only to the specific person, on the specific device, for the specific job each time. In addition to the thorough authentication process, BeyondCorp continues to monitor device metadata during use as part of the system’s decision to continue to trust (or not trust) a user. With this information, if a user accidentally exposes the system to malware, for example, access can be revoked quickly. Max and Rob explain the steps Google went through to create such a state-of-the-art security program and give tips on how companies can build something similar. Codifying your employees’ needs and preferences, detailing the levels of

Priyanka Vergadia hops back into the host seat this week, joining Mark Mirchandani to talk to Marit Rødevand of Strise. Strise is an AI-driven enterprise company using knowledge graphs to gather and analyze massive amounts of information, depositing it into a web-based interface to help large clients such as banks solve data-driven problems. Strise’s web-based data platform is customizable, flexible, and capable of keeping up with the fast-paced world of technology so enterprise companies aren’t constantly putting time and resources into reworking old or building new software. To do this, Strise uses knowledge graphs rather than typical databases to create what Marit calls a future-proof data model. Marit explains knowledge graphs in detail, emphasizing that they can reduce training of machine models, allow new data to be input easily, and make analyzing unstructured data much easier. Knowledge graphs take data that would normally only make sense to humans and in effect translate it for computers. Using banki

Josh Brown, Developer Advocate at Spotify, is on the podcast this week with your hosts Mark and Brian. Working in the Open Developer Platform department, Josh supports third-party developers as they create music experiences for users using the Spotify APIs and SDKs. The most popular of these, the Spotify Web API, lets developers access metadata about music and facilitates library management for users. We talk later in the episode about the types of applications developers are creating using the Spotify Web API and how it’s changing the way people listen. Using developer feedback, Spotify has continued to improve on the API, now offering podcast support, for example. With the new podcast support, hobby developers especially are developing apps that make podcast listening easier and more social. To create these open platform APIs, Josh tells us they relied heavily on Google Cloud products like GKE and Cloud Storage. To manage the GCP products they use, Spotify created an internal portal called Backstage. Indepe

It’s cyber security week on the podcast as Priyanka Vergadia joins Mark Mirchandani to talk with the folks of the Chronicle Security Team. Our guests Ansh Patniak and Dr. Anton Chuvakin start the show off with a brief explanation of Chronicle, which is a security analytics platform that can identify threats and correct them. Anton details the threats facing clients today and why it’s important to continue to guard against old threats as well. Cyber security developers must constantly examine the landscape, adjust tools used, and think ahead to try to predict possible future problems. Ansh elaborates, pointing out that sometimes, all the security needed to protect against old, current, and potentially new threats can create a data overload that causes some threats to be lost in a jungle of notifications. Analyzing this data to gain insights about the health of a company’s cyber security is an important part of the process, and Chronicle can help with that. We discuss other challenges in the security analytics

Our guests Justin Lerma and Pathik Sharma join Brian and Mark this week to talk cost optimization techniques for internet projects. Justin and Pathik, both of the Professional Services Organization, work to help customers get the most out of GCP while maintaining their project budgets. They help customers take business success metrics and track them from a cost perspective, allowing the client to get an understanding of how much each business goal actually costs, rather than an aggregate of how much has been spent in total. This information is used to tailor GCP product usage and cost optimization to each client project. Pathik explains how the Recommender API can help with VM usage by suggesting shrinking or removing a VM altogether for cost savings. With thorough analysis, clients can also benefit from cost savings by paying for longterm usage of GCP products rather than month-to-month. For storage and analysis, BigQuery can offer better performance at a lower cost with partitioning and clustering. Througho