Downloaded 32 times
Uploaded byJosh Lane
Azure Cosmos DB - NoSQL In the Microsoft Cloud
This document provides an overview of Azure Cosmos DB, a globally distributed multi-model database service by Microsoft. It discusses how Azure Cosmos DB is purpose-built to handle data at scale across any application, with predictable low latency and high availability. The document also notes some areas for improvement like pricing models and migration quirks, while outlining Azure Cosmos DB's flexibility and potential for future innovations.
More Related Content
Similar to Azure Cosmos DB - NoSQL In the Microsoft Cloud
Azure Cosmos DB - NoSQL In the Microsoft Cloud
- 1. Azure Cosmos DB TheGreat NoSQL Balancing Act in the Microsoft Cloud { “speaker”: “Josh Lane”, “title”: “Azure Content Lead”, “company”: “cloudacademy.com”, “twitter”: “@jplane” }
- 2.
- 3. Oh, what tangled webs weweave… Scale Up Scale Out Stored Procedures Stored Procedures Caching
- 4. Is Your DatabaseAn Unquestioned Assumption?
- 5. How About ThisInstead? Acknowledge Choice is good… but all choices represent some compromise Choose! Make a deliberate choice Goals Meet requirements, but also minimize compromises
- 6. Purpose-built to solvereal-world challenges of data at scale 2010 – “Project Florence” Elastic, global scale of throughput and storage Predictable, low-latency reads and writes Resource-governed architecture Highly available and SLA-backed Minimize/eliminate infrastructure burdens Developer-friendly… multiple data models, APIs, consistency options Azure Cosmos DB
- 7.
- 8. Low-latency Geo-scale Familiar and approachable Flexibleand evolving Cost-efficient The Great Balancing Act
- 10.
- 11.
- 12. Azure Table Storagecompatibility
- 13.
- 14.
- 15. Availability Single-click replication intoany/all of 38+ datacenters Failover Automated and prioritized Manual (for DR testing, “follow the clock”, etc.) Multiple replicas within each configured region Logical or physical API endpoints Logical – transparent multi-homing Physical – prioritized control over connectivity Automated backups
- 16. Perf and Scale Lowlatency reads (10 ms) and writes (15 ms) SLA-backed Topology-agnostic Provisioned, reserved throughput model Request Units per second (or minute) You can change this as needed! Transparent partitioning Single write replica, multi read replicas
- 17.
- 18. Areas for Improvement Pricingmodel “Something something great power something great responsibility…” Migration quirks Triggers are opt-in SQL dialect is a subset API support must be declared up-front Partitions are “transparent”, not transparent Change feed fidelity Portal tooling is… v1
- 19. Designed for Happiness ACIDtransactions (within a partition) Automated and configurable indexing Change feed Geospatial data types Local emulator JavaScript sprocs, triggers, UDFs Default encryption at rest + in transit Configurable firewall Configurable TTL Azure Function binding ODBC driver LINQ (.NET)
- 20. Baseless Speculation on the Future •More and richer query APIs • More and better Azure ecosystem integration • Architectural building blocks • CQRS • Event sourcing • Stateful services • More runtime integration • .NET Core • Azure Functions runtime • Pricing innovation • Spot market • Secondary market • ML-based resource governance • Convergence with SQL Database
- 21.
Editor's Notes
- #3 Let’s begin with this basic assertion… probably not controversial. Let’s also agree that while this may be true, the degree of compromise can differ widely, between competing options.
- #4 Consider the range of data-oriented choices we face on a typical web project…
- #5 Consider the things we do in the name of a better user experience…
- #6 More broadly, beyond feature details… are we simply falling into our choice of database implicitly? Or are we making an explicit decision based on project and team fit?
- #10 In this vein, and with this framework of thought, that I introduce Azure Cosmos DB…
- #12 I like to refer to the work done by the Cosmos DB team as “the great balancing act”, since Cosmos DB is an ever-evolving service (will adapt and grow in capability over time) DocDB customers became Cosmos DB customers overnite
- #13 Further evidence of the deliberate thought and effort made by the Cosmos DB team Not merely making data available at broad scale, but making guaranteed, SLA-backed access to that data available at broad and configurable scale.
- #19 Choose on an account- or read-operation basis
- #24 Great news… virtually all of this is value-add above the core service offering, and can be improved upon over time Pricing model: RU/m is a recent addition