Summary The rapid growth of machine learning, especially large language models, have led to a commensurate growth in the need to store and compare vectors. In this episode Louis Brandy discusses the applications for vector search capabilities both in and outside of AI, as well as the challenges of maintaining real-time indexes of vector data. Announcements Hello and welcome to the Data Engineering Podcast, the show about modern data management Introducing RudderStack Profiles. RudderStack Profiles takes the SaaS guesswork and SQL grunt work out of building complete customer profiles so you can quickly ship actionable, enriched data to every downstream team. You specify the customer traits, then Profiles runs the joins and computations for you to create complete customer profiles. Get all of the details and try the new product today at dataengineeringpodcast.com/rudderstack (https://www.dataengineeringpodcast.com/rudderstack) This episode is brought to you by Datafold – a testing automation platform for data engineers that finds data quality issues before the code and data are deployed to production. Datafold leverages data-diffing to compare production and development environments and column-level lineage to show you the exact impact of every code change on data, metrics, and BI tools, keeping your team productive and stakeholders happy. Datafold integrates with dbt, the modern data stack, and seamlessly plugs in your data CI for team-wide and automated testing. If you are migrating to a modern data stack, Datafold can also help you automate data and code validation to speed up the migration. Learn more about Datafold by visiting dataengineeringpodcast.com/datafold (https://www.dataengineeringpodcast.com/datafold) You shouldn't have to throw away the database to build with fast-changing data. You should be able to keep the familiarity of SQL and the proven architecture of cloud warehouses, but swap the decades-old batch computation model for an efficient incremental engine to get complex queries that are always up-to-date. With Materialize, you can! It’s the only true SQL streaming database built from the ground up to meet the needs of modern data products. Whether it’s real-time dashboarding and analytics, personalization and segmentation or automation and alerting, Materialize gives you the ability to work with fresh, correct, and scalable results — all in a familiar SQL interface. Go to dataengineeringpodcast.com/materialize (https://www.dataengineeringpodcast.com/materialize) today to get 2 weeks free! If you’re a data person, you probably have to jump between different tools to run queries, build visualizations, write Python, and send around a lot of spreadsheets and CSV files. Hex brings everything together. Its powerful notebook UI lets you analyze data in SQL, Python, or no-code, in any combination, and work together with live multiplayer and version control. And now, Hex’s magical AI tools can generate queries and code, create visualizations, and even kickstart a whole analysis for you – all from natural language prompts. It’s like having an analytics co-pilot built right into where you’re already doing your work. Then, when you’re ready to share, you can use Hex’s drag-and-drop app builder to configure beautiful reports or dashboards that anyone can use. Join the hundreds of data teams like Notion, AllTrails, Loom, Mixpanel and Algolia using Hex every day to make their work more impactful. Sign up today at dataengineeringpodcast.com/hex (https://www.dataengineeringpodcast.com/hex) to get a 30-day free trial of the Hex Team plan! Your host is Tobias Macey and today I'm interviewing Louis Brandy about building vector indexes in real-time for analytics and AI applications Interview Introduction How did you get involved in the area of data management? Can you describe what vector search is and how it differs from other search technologies? What are the technical challenges related to providing vector search? What are the applications for vector search that merit the added complexity? Vector databases have been gaining a lot of attention recently with the proliferation of LLM applications. Is a dedicated database technology required to support vector indexes/vector search queries? What are the use cases for native vector data types that are separate from AI? With the increasing usage of vectors for data and AI/ML applications, who do you typically see as the owner of that problem space? (e.g. data engineers, ML engineers, data scientists, etc.) For teams who are investing in vector search, what are the architectural considerations that they need to be aware of? How does it impact the data pipeline strategies/topologies used? What are the complexities that need to be addressed when updating vector data in a real-time/streaming fashion? How does that influence the client strategies that are querying that data? What are the most interesting, innovative, or unexpected ways that you have seen vector search used? What are the most interesting, unexpected, or challenging lessons that you have learned while working on vector search applications? When is vector search the wrong choice? What do you see as future potential applications for vector indexes/vector search? Contact Info LinkedIn (https://www.linkedin.com/in/lbrandy/) Parting Question From your perspective, what is the biggest gap in the tooling or technology for data management today? Closing Announcements Thank you for listening! Don't forget to check out our other shows. The Machine Learning Podcast (https://www.themachinelearningpodcast.com) helps you go from idea to production with machine learning. Podcast.__init__ (https://www.pythonpodcast.com) covers the Python language, its community, and the innovative ways it is being used. Visit the site (https://www.dataengineeringpodcast.com) to subscribe to the show, sign up for the mailing list, and read the show notes. If you've learned something or tried out a project from the show then tell us about it! Email [email protected] (mailto:[email protected])) with your story. To help other people find the show please leave a review on Apple Podcasts (https://podcasts.apple.com/us/podcast/data-engineering-podcast/id1193040557) and tell your friends and co-workers Links Rockset (https://rockset.com/) Podcast Episode (https://www.dataengineeringpodcast.com/rockset-serverless-analytics-episode-101/) Vector Index (https://www.datastax.com/guides/what-is-a-vector-index) Vector Search (https://www.datastax.com/guides/what-is-vector-search) Rockset Implementation Explanation (https://rockset.com/videos/vector-search-architecture/) Vector Space (https://en.wikipedia.org/wiki/Vector_space) Euclidean Distance (https://en.wikipedia.org/wiki/Euclidean_distance) OLAP == Online Analytical Processing (https://en.wikipedia.org/wiki/Online_analytical_processing) OLTP == Online Transaction Processing (https://en.wikipedia.org/wiki/Online_transaction_processing) The intro and outro music is from The Hug (http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/Love_death_and_a_drunken_monkey/04_-_The_Hug) by The Freak Fandango Orchestra (http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/) / CC BY-SA (http://creativecommons.org/licenses/by-sa/3.0/)

Summary A significant amount of time in data engineering is dedicated to building connections and semantic meaning around pieces of information. Linked data technologies provide a means of tightly coupling metadata with raw information. In this episode Brian Platz explains how JSON-LD can be used as a shared representation of linked data for building semantic data products. Announcements Hello and welcome to the Data Engineering Podcast, the show about modern data management This episode is brought to you by Datafold – a testing automation platform for data engineers that finds data quality issues before the code and data are deployed to production. Datafold leverages data-diffing to compare production and development environments and column-level lineage to show you the exact impact of every code change on data, metrics, and BI tools, keeping your team productive and stakeholders happy. Datafold integrates with dbt, the modern data stack, and seamlessly plugs in your data CI for team-wide and automated testing. If you are migrating to a modern data stack, Datafold can also help you automate data and code validation to speed up the migration. Learn more about Datafold by visiting dataengineeringpodcast.com/datafold (https://www.dataengineeringpodcast.com/datafold) Introducing RudderStack Profiles. RudderStack Profiles takes the SaaS guesswork and SQL grunt work out of building complete customer profiles so you can quickly ship actionable, enriched data to every downstream team. You specify the customer traits, then Profiles runs the joins and computations for you to create complete customer profiles. Get all of the details and try the new product today at dataengineeringpodcast.com/rudderstack (https://www.dataengineeringpodcast.com/rudderstack) You shouldn't have to throw away the database to build with fast-changing data. You should be able to keep the familiarity of SQL and the proven architecture of cloud warehouses, but swap the decades-old batch computation model for an efficient incremental engine to get complex queries that are always up-to-date. With Materialize, you can! It’s the only true SQL streaming database built from the ground up to meet the needs of modern data products. Whether it’s real-time dashboarding and analytics, personalization and segmentation or automation and alerting, Materialize gives you the ability to work with fresh, correct, and scalable results — all in a familiar SQL interface. Go to dataengineeringpodcast.com/materialize (https://www.dataengineeringpodcast.com/materialize) today to get 2 weeks free! If you’re a data person, you probably have to jump between different tools to run queries, build visualizations, write Python, and send around a lot of spreadsheets and CSV files. Hex brings everything together. Its powerful notebook UI lets you analyze data in SQL, Python, or no-code, in any combination, and work together with live multiplayer and version control. And now, Hex’s magical AI tools can generate queries and code, create visualizations, and even kickstart a whole analysis for you – all from natural language prompts. It’s like having an analytics co-pilot built right into where you’re already doing your work. Then, when you’re ready to share, you can use Hex’s drag-and-drop app builder to configure beautiful reports or dashboards that anyone can use. Join the hundreds of data teams like Notion, AllTrails, Loom, Mixpanel and Algolia using Hex every day to make their work more impactful. Sign up today at dataengineeringpodcast.com/hex (https://www.dataengineeringpodcast.com/hex) to get a 30-day free trial of the Hex Team plan! Your host is Tobias Macey and today I'm interviewing Brian Platz about using JSON-LD for building linked-data products Interview Introduction How did you get involved in the area of data management? Can you describe what the term "linked data product" means and some examples of when you might build one? What is the overlap between knowledge graphs and "linked data products"? What is JSON-LD? What are the domains in which it is typically used? How does it assist in developing linked data products? what are the characteristics that distinguish a knowledge graph from What are the layers/stages of applications and data that can/should incorporate JSON-LD as the representation for records and events? What is the level of native support/compatibiliity that you see for JSON-LD in data systems? What are the modeling exercises that are necessary to ensure useful and appropriate linkages of different records within and between products and organizations? Can you describe the workflow for building autonomous linkages across data assets that are modelled as JSON-LD? What are the most interesting, innovative, or unexpected ways that you have seen JSON-LD used for data workflows? What are the most interesting, unexpected, or challenging lessons that you have learned while working on linked data products? When is JSON-LD the wrong choice? What are the future directions that you would like to see for JSON-LD and linked data in the data ecosystem? Contact Info LinkedIn (https://www.linkedin.com/in/brianplatz/) Parting Question From your perspective, what is the biggest gap in the tooling or technology for data management today? Closing Announcements Thank you for listening! Don't forget to check out our other shows. Podcast.__init__ (https://www.pythonpodcast.com) covers the Python language, its community, and the innovative ways it is being used. The Machine Learning Podcast (https://www.themachinelearningpodcast.com) helps you go from idea to production with machine learning. Visit the site (https://www.dataengineeringpodcast.com) to subscribe to the show, sign up for the mailing list, and read the show notes. If you've learned something or tried out a project from the show then tell us about it! Email [email protected] (mailto:[email protected])) with your story. To help other people find the show please leave a review on Apple Podcasts (https://podcasts.apple.com/us/podcast/data-engineering-podcast/id1193040557) and tell your friends and co-workers Links Fluree (https://flur.ee/) JSON-LD (https://json-ld.org/) Knowledge Graph (https://en.wikipedia.org/wiki/Knowledge_graph) Adjacency List (https://en.wikipedia.org/wiki/Adjacency_list) RDF == Resource Description Framework (https://www.w3.org/RDF/) Semantic Web (https://en.wikipedia.org/wiki/Semantic_Web) Open Graph (https://ogp.me/) Schema.org (https://schema.org/) RDF Triple (https://en.wikipedia.org/wiki/Semantic_triple) IDMP == Identification of Medicinal Products (https://www.fda.gov/industry/fda-data-standards-advisory-board/identification-medicinal-products-idmp) FIBO == Financial Industry Business Ontology (https://spec.edmcouncil.org/fibo/) OWL Standard (https://www.w3.org/OWL/) NP-Hard (https://en.wikipedia.org/wiki/NP-hardness) Forward-Chaining Rules (https://en.wikipedia.org/wiki/Forward_chaining) SHACL == Shapes Constraint Language) (https://www.w3.org/TR/shacl/) Zero Knowledge Cryptography (https://en.wikipedia.org/wiki/Zero-knowledge_proof) Turtle Serialization (https://www.w3.org/TR/turtle/) The intro and outro music is from The Hug (http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/Love_death_and_a_drunken_monkey/04_-_The_Hug) by The Freak Fandango Orchestra (http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/) / CC BY-SA (http://creativecommons.org/licenses/by-sa/3.0/)

Summary Data systems are inherently complex and often require integration of multiple technologies. Orchestrators are centralized utilities that control the execution and sequencing of interdependent operations. This offers a single location for managing visibility and error handling so that data platform engineers can manage complexity. In this episode Nick Schrock, creator of Dagster, shares his perspective on the state of data orchestration technology and its application to help inform its implementation in your environment. Announcements Hello and welcome to the Data Engineering Podcast, the show about modern data management Introducing RudderStack Profiles. RudderStack Profiles takes the SaaS guesswork and SQL grunt work out of building complete customer profiles so you can quickly ship actionable, enriched data to every downstream team. You specify the customer traits, then Profiles runs the joins and computations for you to create complete customer profiles. Get all of the details and try the new product today at dataengineeringpodcast.com/rudderstack (https://www.dataengineeringpodcast.com/rudderstack) This episode is brought to you by Datafold – a testing automation platform for data engineers that finds data quality issues before the code and data are deployed to production. Datafold leverages data-diffing to compare production and development environments and column-level lineage to show you the exact impact of every code change on data, metrics, and BI tools, keeping your team productive and stakeholders happy. Datafold integrates with dbt, the modern data stack, and seamlessly plugs in your data CI for team-wide and automated testing. If you are migrating to a modern data stack, Datafold can also help you automate data and code validation to speed up the migration. Learn more about Datafold by visiting dataengineeringpodcast.com/datafold (https://www.dataengineeringpodcast.com/datafold) You shouldn't have to throw away the database to build with fast-changing data. You should be able to keep the familiarity of SQL and the proven architecture of cloud warehouses, but swap the decades-old batch computation model for an efficient incremental engine to get complex queries that are always up-to-date. With Materialize, you can! It’s the only true SQL streaming database built from the ground up to meet the needs of modern data products. Whether it’s real-time dashboarding and analytics, personalization and segmentation or automation and alerting, Materialize gives you the ability to work with fresh, correct, and scalable results — all in a familiar SQL interface. Go to dataengineeringpodcast.com/materialize (https://www.dataengineeringpodcast.com/materialize) today to get 2 weeks free! Your host is Tobias Macey and today I'm welcoming back Nick Schrock to talk about the state of the ecosystem for data orchestration Interview Introduction How did you get involved in the area of data management? Can you start by defining what data orchestration is and how it differs from other types of orchestration systems? (e.g. container orchestration, generalized workflow orchestration, etc.) What are the misconceptions about the applications of/need for/cost to implement data orchestration? How do those challenges of customer education change across roles/personas? Because of the multi-faceted nature of data in an organization, how does that influence the capabilities and interfaces that are needed in an orchestration engine? You have been working on Dagster for five years now. How have the requirements/adoption/application for orchestrators changed in that time? One of the challenges for any orchestration engine is to balance the need for robust and extensible core capabilities with a rich suite of integrations to the broader data ecosystem. What are the factors that you have seen make the most influence in driving adoption of a given engine? What are the most interesting, innovative, or unexpected ways that you have seen data orchestration implemented and/or used? What are the most interesting, unexpected, or challenging lessons that you have learned while working on data orchestration? When is a data orchestrator the wrong choice? What do you have planned for the future of orchestration with Dagster? Contact Info @schrockn (https://twitter.com/schrockn) on Twitter LinkedIn (https://www.linkedin.com/in/schrockn) Parting Question From your perspective, what is the biggest gap in the tooling or technology for data management today? Closing Announcements Thank you for listening! Don't forget to check out our other shows. Podcast.__init__ (https://www.pythonpodcast.com) covers the Python language, its community, and the innovative ways it is being used. The Machine Learning Podcast (https://www.themachinelearningpodcast.com) helps you go from idea to production with machine learning. Visit the site (https://www.dataengineeringpodcast.com) to subscribe to the show, sign up for the mailing list, and read the show notes. If you've learned something or tried out a project from the show then tell us about it! Email [email protected] (mailto:[email protected])) with your story. To help other people find the show please leave a review on Apple Podcasts (https://podcasts.apple.com/us/podcast/data-engineering-podcast/id1193040557) and tell your friends and co-workers Links Dagster (https://dagster.io/) GraphQL (https://graphql.org/) K8s == Kubernetes (https://kubernetes.io/) Airbyte (https://airbyte.com/) Podcast Episode (https://www.dataengineeringpodcast.com/airbyte-open-source-data-integration-episode-173/) Hightouch (https://hightouch.com/) Podcast Episode (https://www.dataengineeringpodcast.com/hightouch-customer-data-warehouse-episode-168/) Airflow (https://airflow.apache.org/) Prefect (https://www.prefect.io) Flyte (https://flyte.org/) Podcast Episode (https://www.dataengineeringpodcast.com/flyte-data-orchestration-machine-learning-episode-291/) dbt (https://www.getdbt.com/) Podcast Episode (https://www.dataengineeringpodcast.com/dbt-data-analytics-episode-81/) DAG == Directed Acyclic Graph (https://en.wikipedia.org/wiki/Directed_acyclic_graph) Temporal (https://temporal.io/) Software Defined Assets (https://docs.dagster.io/concepts/assets/software-defined-assets) DataForm (https://dataform.co/) Gradient Flow State Of Orchestration Report 2022 (https://gradientflow.com/2022-workflow-orchestration-survey/) MLOps Is 98% Data Engineering (https://mlops.community/mlops-is-mostly-data-engineering/) DataHub (https://datahubproject.io/) Podcast Episode (https://www.dataengineeringpodcast.com/datahub-metadata-management-episode-147/) OpenMetadata (https://open-metadata.org/) Podcast Episode (https://www.dataengineeringpodcast.com/openmetadata-universal-metadata-layer-episode-237/) Atlan (https://atlan.com/) Podcast Episode (https://www.dataengineeringpodcast.com/atlan-data-team-collaboration-episode-179/) The intro and outro music is from The Hug (http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/Love_death_and_a_drunken_monkey/04_-_The_Hug) by The Freak Fandango Orchestra (http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/) / CC BY-SA (http://creativecommons.org/licenses/by-sa/3.0/)

Summary Cloud data warehouses and the introduction of the ELT paradigm has led to the creation of multiple options for flexible data integration, with a roughly equal distribution of commercial and open source options. The challenge is that most of those options are complex to operate and exist in their own silo. The dlt project was created to eliminate overhead and bring data integration into your full control as a library component of your overall data system. In this episode Adrian Brudaru explains how it works, the benefits that it provides over other data integration solutions, and how you can start building pipelines today. Announcements Hello and welcome to the Data Engineering Podcast, the show about modern data management Introducing RudderStack Profiles. RudderStack Profiles takes the SaaS guesswork and SQL grunt work out of building complete customer profiles so you can quickly ship actionable, enriched data to every downstream team. You specify the customer traits, then Profiles runs the joins and computations for you to create complete customer profiles. Get all of the details and try the new product today at dataengineeringpodcast.com/rudderstack (https://www.dataengineeringpodcast.com/rudderstack) You shouldn't have to throw away the database to build with fast-changing data. You should be able to keep the familiarity of SQL and the proven architecture of cloud warehouses, but swap the decades-old batch computation model for an efficient incremental engine to get complex queries that are always up-to-date. With Materialize, you can! It’s the only true SQL streaming database built from the ground up to meet the needs of modern data products. Whether it’s real-time dashboarding and analytics, personalization and segmentation or automation and alerting, Materialize gives you the ability to work with fresh, correct, and scalable results — all in a familiar SQL interface. Go to dataengineeringpodcast.com/materialize (https://www.dataengineeringpodcast.com/materialize) today to get 2 weeks free! This episode is brought to you by Datafold – a testing automation platform for data engineers that finds data quality issues before the code and data are deployed to production. Datafold leverages data-diffing to compare production and development environments and column-level lineage to show you the exact impact of every code change on data, metrics, and BI tools, keeping your team productive and stakeholders happy. Datafold integrates with dbt, the modern data stack, and seamlessly plugs in your data CI for team-wide and automated testing. If you are migrating to a modern data stack, Datafold can also help you automate data and code validation to speed up the migration. Learn more about Datafold by visiting dataengineeringpodcast.com/datafold (https://www.dataengineeringpodcast.com/datafold) Your host is Tobias Macey and today I'm interviewing Adrian Brudaru about dlt, an open source python library for data loading Interview Introduction How did you get involved in the area of data management? Can you describe what dlt is and the story behind it? What is the problem you want to solve with dlt? Who is the target audience? The obvious comparison is with systems like Singer/Meltano/Airbyte in the open source space, or Fivetran/Matillion/etc. in the commercial space. What are the complexities or limitations of those tools that leave an opening for dlt? Can you describe how dlt is implemented? What are the benefits of building it in Python? How have the design and goals of the project changed since you first started working on it? How does that language choice influence the performance and scaling characteristics? What problems do users solve with dlt? What are the interfaces available for extending/customizing/integrating with dlt? Can you talk through the process of adding a new source/destination? What is the workflow for someone building a pipeline with dlt? How does the experience scale when supporting multiple connections? Given the limited scope of extract and load, and the composable design of dlt it seems like a purpose built companion to dbt (down to the naming). What are the benefits of using those tools in combination? What are the most interesting, innovative, or unexpected ways that you have seen dlt used? What are the most interesting, unexpected, or challenging lessons that you have learned while working on dlt? When is dlt the wrong choice? What do you have planned for the future of dlt? Contact Info LinkedIn (https://www.linkedin.com/in/data-team/?originalSubdomain=de) Join our community to discuss further (https://join.slack.com/t/dlthub-community/shared_invite/zt-1slox199h-HAE7EQoXmstkP_bTqal65g) Parting Question From your perspective, what is the biggest gap in the tooling or technology for data management today? Closing Announcements Thank you for listening! Don't forget to check out our other shows. Podcast.__init__ (https://www.pythonpodcast.com) covers the Python language, its community, and the innovative ways it is being used. The Machine Learning Podcast (https://www.themachinelearningpodcast.com) helps you go from idea to production with machine learning. Visit the site (https://www.dataengineeringpodcast.com) to subscribe to the show, sign up for the mailing list, and read the show notes. If you've learned something or tried out a project from the show then tell us about it! Email [email protected] (mailto:[email protected])) with your story. To help other people find the show please leave a review on Apple Podcasts (https://podcasts.apple.com/us/podcast/data-engineering-podcast/id1193040557) and tell your friends and co-workers Links dlt (https://dlthub.com/) Harness Success Story (https://dlthub.com/success-stories/harness/) Our guiding product principles (https://dlthub.com/product/) Ecosystem support (https://dlthub.com/docs/dlt-ecosystem) From basic to complex, dlt has many capabilities (https://dlthub.com/docs/getting-started/build-a-data-pipeline) Singer (https://www.singer.io/) Airbyte (https://airbyte.com/) Podcast Episode (https://www.dataengineeringpodcast.com/airbyte-open-source-data-integration-episode-173/) Meltano (https://meltano.com/) Podcast Episode (https://www.dataengineeringpodcast.com/meltano-data-integration-episode-141/) Matillion (https://www.matillion.com/) Podcast Episode (https://www.dataengineeringpodcast.com/matillion-cloud-data-integration-episode-286/) Fivetran (https://www.fivetran.com/) Podcast Episode (https://www.dataengineeringpodcast.com/fivetran-data-replication-episode-93/) DuckDB (https://duckdb.org/) Podcast Episode (https://www.dataengineeringpodcast.com/duckdb-in-process-olap-database-episode-270/) OpenAPI (https://www.openapis.org/) Data Mesh (https://martinfowler.com/articles/data-monolith-to-mesh.html) Podcast Episode (https://www.dataengineeringpodcast.com/data-mesh-revisited-episode-250/) SQLMesh (https://sqlmesh.com/) Podcast Episode (https://www.dataengineeringpodcast.com/sqlmesh-open-source-dataops-episode-380) Airflow (https://airflow.apache.org/) Dagster (https://dagster.io/) Podcast Episode (https://www.dataengineeringpodcast.com/dagster-data-platform-big-complexity-episode-239/) Prefect (https://www.prefect.io/) Podcast Episode (https://www.dataengineeringpodcast.com/prefect-workflow-engine-episode-86/) Alto (https://github.com/z3z1ma/alto) The intro and outro music is from The Hug (http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/Love_death_and_a_drunken_monkey/04_-_The_Hug) by The Freak Fandango Orchestra (http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/) / CC BY-SA (http://creativecommons.org/licenses/by-sa/3.0/)

Summary Data persistence is one of the most challenging aspects of computer systems. In the era of the cloud most developers rely on hosted services to manage their databases, but what if you are a cloud service? In this episode Vignesh Ravichandran explains how his team at Cloudflare provides PostgreSQL as a service to their developers for low latency and high uptime services at global scale. This is an interesting and insightful look at pragmatic engineering for reliability and scale. Announcements Hello and welcome to the Data Engineering Podcast, the show about modern data management Introducing RudderStack Profiles. RudderStack Profiles takes the SaaS guesswork and SQL grunt work out of building complete customer profiles so you can quickly ship actionable, enriched data to every downstream team. You specify the customer traits, then Profiles runs the joins and computations for you to create complete customer profiles. Get all of the details and try the new product today at dataengineeringpodcast.com/rudderstack (https://www.dataengineeringpodcast.com/rudderstack) This episode is brought to you by Datafold – a testing automation platform for data engineers that finds data quality issues before the code and data are deployed to production. Datafold leverages data-diffing to compare production and development environments and column-level lineage to show you the exact impact of every code change on data, metrics, and BI tools, keeping your team productive and stakeholders happy. Datafold integrates with dbt, the modern data stack, and seamlessly plugs in your data CI for team-wide and automated testing. If you are migrating to a modern data stack, Datafold can also help you automate data and code validation to speed up the migration. Learn more about Datafold by visiting dataengineeringpodcast.com/datafold (https://www.dataengineeringpodcast.com/datafold) You shouldn't have to throw away the database to build with fast-changing data. You should be able to keep the familiarity of SQL and the proven architecture of cloud warehouses, but swap the decades-old batch computation model for an efficient incremental engine to get complex queries that are always up-to-date. With Materialize, you can! It’s the only true SQL streaming database built from the ground up to meet the needs of modern data products. Whether it’s real-time dashboarding and analytics, personalization and segmentation or automation and alerting, Materialize gives you the ability to work with fresh, correct, and scalable results — all in a familiar SQL interface. Go to dataengineeringpodcast.com/materialize (https://www.dataengineeringpodcast.com/materialize) today to get 2 weeks free! Your host is Tobias Macey and today I'm interviewing Vignesh Ravichandran about building an internal database as a service platform at Cloudflare Interview Introduction How did you get involved in the area of data management? Can you start by describing the different database workloads that you have at Cloudflare? What are the different methods that you have used for managing database instances? What are the requirements and constraints that you had to account for in designing your current system? Why Postgres? optimizations for Postgres simplification from not supporting multiple engines limitations in postgres that make multi-tenancy challenging scale of operation (data volume, request rate What are the most interesting, innovative, or unexpected ways that you have seen your DBaaS used? What are the most interesting, unexpected, or challenging lessons that you have learned while working on your internal database platform? When is an internal database as a service the wrong choice? What do you have planned for the future of Postgres hosting at Cloudflare? Contact Info LinkedIn (https://www.linkedin.com/in/vigneshravichandran28/) Website (https://viggy28.dev/) Parting Question From your perspective, what is the biggest gap in the tooling or technology for data management today? Closing Announcements Thank you for listening! Don't forget to check out our other shows. Podcast.__init__ (https://www.pythonpodcast.com) covers the Python language, its community, and the innovative ways it is being used. The Machine Learning Podcast (https://www.themachinelearningpodcast.com) helps you go from idea to production with machine learning. Visit the site (https://www.dataengineeringpodcast.com) to subscribe to the show, sign up for the mailing list, and read the show notes. If you've learned something or tried out a project from the show then tell us about it! Email [email protected] (mailto:[email protected])) with your story. To help other people find the show please leave a review on Apple Podcasts (https://podcasts.apple.com/us/podcast/data-engineering-podcast/id1193040557) and tell your friends and co-workers Links Cloudflare (https://www.cloudflare.com/) PostgreSQL (https://www.postgresql.org/) Podcast Episode (https://www.dataengineeringpodcast.com/postgresql-with-jonathan-katz-episode-42/) IP Address Data Type in Postgres (https://www.postgresql.org/docs/current/datatype-net-types.html) CockroachDB (https://www.cockroachlabs.com/) Podcast Episode (https://www.dataengineeringpodcast.com/cockroachdb-with-peter-mattis-episode-35/) Citus (https://www.citusdata.com/) Podcast Episode (https://www.dataengineeringpodcast.com/citus-data-with-ozgun-erdogan-and-craig-kerstiens-episode-13/) Yugabyte (https://www.yugabyte.com/) Podcast Episode (https://www.dataengineeringpodcast.com/yugabytedb-planet-scale-sql-episode-115/) Stolon (https://github.com/sorintlab/stolon) pg_rewind (https://www.postgresql.org/docs/current/app-pgrewind.html) PGBouncer (https://www.pgbouncer.org/) HAProxy Presentation (https://www.youtube.com/watch?v=HIOo4j-Tiq4) Etcd (https://etcd.io/) Patroni (https://patroni.readthedocs.io/en/latest/) pg_upgrade (https://www.postgresql.org/docs/current/pgupgrade.html) Edge Computing (https://en.wikipedia.org/wiki/Edge_computing) The intro and outro music is from The Hug (http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/Love_death_and_a_drunken_monkey/04_-_The_Hug) by The Freak Fandango Orchestra (http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/) / CC BY-SA (http://creativecommons.org/licenses/by-sa/3.0/)