Databricks Platform.pptx
Download as PPTX, PDF7 likes18,977 views
The document provides an overview of the Databricks platform, which offers a unified environment for data engineering, analytics, and AI. It describes how Databricks addresses the complexity of managing data across siloed systems by providing a single "data lakehouse" platform where all data and analytics workloads can be run. Key features highlighted include Delta Lake for ACID transactions on data lakes, auto loader for streaming data ingestion, notebooks for interactive coding, and governance tools to securely share and catalog data and models.
Databricks Platform.pptx
- 1. The Databricks Platform Introduction All your data, analytics and AI on one platform Alex Ivanichev March 2022
- 2. DataBricks is a unified & open Data and Analytics Platform What is DataBricks ?
- 3. Modern Data Teams 5 Data Engineers Data Scientists Data Analysts
- 4. How the data management looks like today ?
- 5. Data management complexity Siloed stacks increase data architecture complexity Data Warehousing Data Engineering Streaming decrease productivity Data Science and ML Data Analysts Data Engineers Data Engineers Disconnected systems and proprietary data formats make integration difficult Data Scientists Amazon Redshift Azure Synapse Snowflake SAP Teradata Google BigQuery IBM Db2 Oracle Autonomous Data Warehouse Hadoop Apache Airflow Apache Kafka Apache Spark Jupyter Amazon SageMaker Amazon EMR Apache Spark Apache Flink Amazon Kinesis Azure ML Studio MatLAB Google Dataproc Cloudera Azure Stream Analytics Google Dataflow Domino Data Labs SAS Tibco Spotfire Confluent TensorFlow PyTorch Extract Load Transform Real-time Database Analytics and BI Data marts Data prep Machine Learning Data Science Streaming Data Engine Data Lake Data Lake Data warehouse Structured, semi- structured and unstructured data Structured, semi- structured and unstructured data Structured data Streaming data sources 5
- 6. Data Warehouse Data Lake vs.
- 7. Warehouses and lakes create complexity Two separate copies of the data Warehouses Proprietary Lakes Open Incompatible interfaces Warehouses SQL Lakes Python Incompatible security and governance models Warehouses Tables Lakes Files
- 8. Data Warehouse Data Lake Streaming Analytics B I Data Science Machine Learning Structured, Semi-Structured and Unstructured Data Data Lakehouse One platform to unify all of your data, analytics, and AI workloads
- 10. The data lakehouse offers a better path Data processing and management built on open source and open standards Common security, governance, and administration Modern Data Engineering Analytics and Data Warehousing Data Science and ML Integrated and collaborative role-based experiences with open API’s Cloud Data Lake Structured, semi-structured, and unstructured data Lake-first approach that builds upon where the freshest, most complete data resides AI/ML from the ground up High reliability and performance Single approach to managing data Support for all use cases on a single platform: • Data engineering • Data warehousing • Real time streaming • Data science and ML Built on open source and open standards Multi-cloud, work with your cloud of choice
- 12. ©2021 Databricks Inc. — All rights r eserved An open approach to bringing data management and governance to data lakes Better reliability with transactions 48x faster data processing with indexing Data governance at scale with fine- grained access control lists Data Warehouse Data Lake
- 13. What is Delta Lake? ● A open source project that enables building a Lakehouse architecture on top of data lakes. ● An storage layer that brings scalable, ACID transactions to Apache Spark and other big-data engines. ● Delta Lake provides ACID transactions, scalable metadata handling, and unifies streaming and batch data processing on top of existing data lakes, such as S3, ADLS, GCS, and HDFS. ● ACID Transactions ● Scalable Metadata Handling ● Time Travel (data versioning) ● Open Format ● Delta Lake change data feed ● Unified Batch and Streaming Source and Sink ● Schema Enforcement ● Schema Evolution ● Audit History ● Updates and Delete ● 100% Compatible with Apache Spark API ● Data Clean-up Key Features: https://databricks.com/blog/2019/08/21/diving-into-delta-lake-unpacking-the-transaction-log.html
- 14. Delta Lake solves challenges with data lakes RELIABILITY & QUALITY PERFORMANCE & LATENCY GOVERNANCE ACID transactions Advanced indexing & caching Governance with Data Catalogs
- 15. Delta Lake key feature - ACID transaction ● Add File: It adds the data file ● Remove File: It removes the data file ● Update Metadata: It updates the table metadata. ● Set Transaction: It records that a structure streaming job created a micro-batch with ID ● Change Protocol: Makes more secure by transferring Delta Lakes to the latest securing protocol. ● Commit Info: It contains the information about the Commits.
- 16. State Recomputing With Checkpoint Files ● Delta Lake automatically generates checkpoint files every 10 commits ● Delta Lake saves a checkpoint file in Parquet format in the same _delta_log subdirectory.
- 17. Building the foundation of a Lakehouse Filtered, Cleaned, Augmented Business-level Aggregates Greatly improve the quality of your data for end users BRONZE SILVER GOLD Raw Ingestion and History Kinesis CSV, JSON, TXT… Data Lake Quality BI & Reporting Streaming Analytics Data Science & ML
- 18. But the reality is not so simple Maintaining data quality and reliability at scale is complex and brittle CSV, JSON, TXT… Data Lake Kinesis BI & Reporting Streaming Analytics Data Science & ML
- 19. Modern data engineering on the lakehouse Data Engineering on the Databricks Lakehouse Platform Open format storage Data transformation Scheduling & orchestration Automatic deployment & operations BI / Reporting Dashboarding Machine Learning / Data Science Data & ML Sharing Data Products Databases Streaming Sources Cloud Object Stores SaaS Applications NoSQL On-premises Systems Data Sources Data Consumers Observability, lineage, and end-to-end pipeline visibility Data quality management Data ingestion
- 20. Data Science & Engineering Workspace
- 21. Databricks Workspaces: Clusters It is a set of computation resources where a developer can run Data Analytics, Data Science, or Data Engineering workloads. The workloads can be executed in the form of a set of commands written in a notebook
- 22. Databricks Workspaces: Notebooks It is a Web Interface where a developer can write and execute codes. Notebook contains a sequence of runnable cells that helps a developer to work with files, manipulate tables, create visualizations, and add narrative texts
- 23. Databricks Workspaces: AutoLoader Auto Loader incrementally and efficiently processes new data files as they arrive in cloud storage. Auto Loader can load data files from Google Cloud Storage (GCS, gs://) in addition to Databricks File System (DBFS, dbfs:/) ** Supports: JSON, CSV, PARQUET, AVRO, ORC, TEXT, and BINARYFILE file formats. ** val checkpoint_path = "/tmp/delta/population_data/_checkpoints" val write_path = "/tmp/delta/population_data" // Set up the stream to begin reading incoming files from the // upload_path location. val df = spark.readStream.format("cloudFiles") .option("cloudFiles.format", "csv") .option("header", "true") .schema("city string, year int, population long") .load(upload_path) // Start the stream. // Use the checkpoint_path location to keep a record of all files that // have already been uploaded to the upload_path location. // For those that have been uploaded since the last check, // write the newly-uploaded files' data to the write_path location. df.writeStream.format("delta") .option("checkpointLocation", checkpoint_path) .start(write_path) https://docs.databricks.com/spark/latest/structured-streaming/auto-loader.html
- 24. Databricks Workspaces: Jobs Jobs allow a user to run notebooks on a scheduled basis. It is a method of executing or automating specific tasks like ETL, Model Building, and more. The pipeline of the ML workflow can be organized into jobs so that it sequentially runs the series of steps one after another
- 25. Databricks Workspaces:Delta Live Tables Delta Live Tables is a framework designed to enable declaratively define, deploy, test & upgrade data pipelines and eliminate operational burdens associated with the management of such pipelines.
- 26. Databricks Workspaces: Repos To empower the process of ML application development, repo’s provide repository- level integration with Git-based hosting providers such as GitHub, GitLab, bitBucket, and Azure DevOps Developers can write code in a Notebook and Sync it with the hosting provider, allowing developers to clone, manage branches, push changes, pull changes, etc.
- 27. Databricks Workspaces: Models It refers to a Developer’s ML Workflow Model registered in the MLflow Model Registry, a centralized model store that manages the entire life cycle of MLflow models. MLflow Model Registry provides all the information about modern lineage, model versioning, present condition, workflow, and stage transition (whether promoted to production or archived).
- 28. Governance requirements for data are quickly evolving
- 29. Governance is hard to enforce on data lakes 42 Cloud 2 Cloud 3 Structured Semi-structured Unstructured Streaming Cloud 1
- 30. The problem is getting bigger Enterprises need a way to share and govern a wide variety of data products Files Dashboards Models Tables
- 31. Unity Catalog for Lakehouse Governance • Centrally catalog, Search, and discover data and AI assets • Simplify governance with a unified Cross- cloud governance model • Easily integrate with your existing Enterprise Data Catalogs • Securely share live data across platforms with delta sharing
- 32. Delta Sharing on Databricks Delta Lake Table Delta Sharing Server Delta Sharing Protocol Data Provider Data Recipient Any Sharing Client Access permissions
- 34. ML Architecture: Data Warehouse VS Data Lakehouse Data Warehouse Data Lakehouse
- 35. Open Multi-Cloud Data Lakehouse and Feature Store Collaborative Multi-Language Notebooks ← Full ML Lifecycle → Model Tracking and Registry Model Training and Tuning Model Serving and Monitoring Automation and Governance Data Science and Machine Learning A data-native and collaborative solution for the full ML lifecycle
- 36. What Does ML Need from a Lakehouse? 58 Access to Unstructured Data • Images, text, audio, custom formats • Libraries understand files, not tables • Must scale to petabytes Open Source Libraries • OSS dominates ML tooling (Tensorflow, scikit- learn, xgboost, R, etc) • Must be able to apply these in Python, R Specialized Hardware, Distributed Compute • Scalability of algorithms • GPUs, for deep learning • Cloud elasticity to manage that cost! Model Lifecycle Management • Outputs are model artifacts • Artifact lineage • Productionization of model
- 37. Three Data Users • SQL and BI tools • Prepare and run reports • Summarize data • Visualize data • (Sometimes) Big Data • Data Warehouse data store • R, SAS, some Python • Statistical analysis • Explain data • Visualize data • Often small data sets • Database, data warehouse data store; local files Business Intelligence Data Science • Python • Deep learning and specialized GPU hardware • Create predictive models • Deploy models to prod • Often big data sets • Unstructured data in files Machine Learning
- 38. How Is ML Different? • Operates on unstructured data like text and images • Can require learning from massive data sets, not just analysis of a sample • Uses open source tooling to manipulate data as “DataFrames” rather than with SQL • Outputs are models rather than data or reports • Sometimes needs special hardware
- 39. MLOps and the Lakehouse • Applying open tools in-place to data in the lakehouse is a win for training • Applying them for operating models is important too! • "Models are data too" • Need to apply models to data • MLFlow for MLOps on the lakehouse • Track and manage model data, lineage, inputs • Deploy models as lakehouse "services"
- 40. Feature Stores for Model Inputs • Tables are OK for managing model input • Input often structured • Well understood, easy to access • … but not quite enough • Upstream lineage: how were features computed? • Downstream lineage: where is the feature used? • Model caller has to read, feed inputs • How to do (also) access in real time?
- 41. SQL Analytics Workspace Query data lake data using familiar ANSI SQL, and find and share new insights faster with the built-in SQL query editor, alerts, visualizations, and interactive dashboards.
- 42. Databricks Workspaces: Queries Provides a simplified control (which is SQL only) to query the data
- 43. Databricks Workspaces: Dashboards A Databricks SQL dashboard lets you combine visualizations and text boxes that provide context with your data.
- 44. Databricks Workspaces: Alerts Alerts notify you when a field returned by a scheduled query meets a threshold. Alerts complement scheduled queries, but their criteria are checked after every execution.
- 45. Databricks Workspaces: Query History The query history shows SQL queries performed using SQL endpoints.
- 46. Thank you