![Disk1
MyBucket
MyObject
part.1
export-xl/
xl.json
{
“version”: “1.0.1”,
“format”: “xl”,
“stat”: {
“size”: 2286,
“modTime”: “2017-12-02T00:24:20.975968336Z”,
},
“erasure”: {
“algorithm”: “klauspost/reedsolomon/vandermonde”,
“data”: 2,
“parity”: 2,
“blockSize”: 10485760,
“index”: 2,
“distribution”: “[
2,
3,
4,
1
],
“checksum”: [
{
“name”: “part.1”,
“algorithm”: “blake2b”,
“hash”: “c24fa0451fd85a3a482c...b672b7f”
}
]
},
“minio”: {
“release”: “DEVELOPMENT.GOGET”
},
“meta”: {
“content-type”: “application/octet-stream”,
“etag”: “c1d217c52d44c9eab00e81496b2b91b6”
},
“parts”: [
{
“number”: 1,
“name”: “part.1”,
“etag”: “”,
“size”: 2286
}
]
}
Disk2
MyBucket
MyObject
part.1
xl.json
Disk3
MyBucket
MyObject
part.1
xl.json
Disk4
MyBucket
MyObject
part.1
xl.json
Erasure Code
Internals](https://image.slidesharecdn.com/minio-january2020-200128172936/85/MinIO-January-2020-Briefing-38-320.jpg)
MinIO January 2020 Briefing
- 1. 1 High Performance, Cloud Native Object Storage January 2020
- 2. high performance, software defined, distributed object storage private cloud MINIO SERVER MINIO CLIENT MINIO SDK 2 What is MinIO
- 3. 3 What Makes Us Different MinIO is focused on performance. We believe we are the fastest object store in existence. MinIO is cloud native. It is the most Kubernetes-friendly solution available for the private cloud. MinIO is 100% open-source giving us an increasingly dominant position in the enterprise. MinIO is built for scale using the same philosophy as the web scalers. MinIO designed for simplicity. Simplicity scales - across clients, clouds and machines.
- 4. 4 What Makes Us the Standard for the Private Cloud DOCKER PULLS GITHUB STARS SLACK MEMBERS CONTRIBUTORS
- 5. 5 Exceptional Growth DOCKER PULLS DIRECT DOWNLOADSUNIQUE USERS 78% YOY Growth. 202% YOY Growth. 265% YOY Growth.
- 6. 6 Developer Traction Kubernetes: 62,537 Docker: 56,181 Minio: 19,956 Apache/Kafka: 14,763 Apache/Mesos: 4,409 *as of january 20th, 2020
- 7. 7 Understanding MinIO’s Reach More than half of the Fortune 500 runs MinIO in some capacity. Of the Fortune 100, 84 companies run MinIO. In the last month we tracked 19K live instances of MinIO across 10K organizations + SIX continents. MinIO is deployed in 153 countries. North America, Europe dominate, followed by China and other developing economies.
- 10. 10 Across Configurations 62% of all instances are containerized using Docker. Only 3% of all instances are running in pure Gateway mode. NAS, Azure, S3 and GCS make up the top four. Linux dominates (81%) followed by Windows (14%) DOCKER GATEWAY OS 27% of all instances managed using Kubernetes - 43% of the containerized instances. KUBERNETES
- 11. 11 MinIO Powers Dozens of Product Offerings MinIO was selected to power the Digital Stream Processing product. Additional products in the pipeline. McKesson’s radiology product uses MinIO for image storage. Datera’s object storage layer is MinIO. MinIO powers the object storage component of this industry leading EHR. “Minio object storage perfectly compliments Portworx by providing a simple object storage service layer on top of Portworx data services.” Nutanix Objects/Buckets is based entirely on MinIO and is referenced in their documentation. The recommended object store for Cloud Private/Kubernetes. “For customers who want to interact with Qumulo via the S3 SDK or API, we recommend using Minio.”
- 12. 12
- 13. 16 Node HDD 25 Gbe Network Avg Read Throughput (GET) Avg Write Throughput (PUT) Distributed 10.81 GB/s 8.57 GB/s Distributed with Encryption 9.38 GB/s 6.91 GB/s S3 Benchmark - Spinning Disk: 16 Nodes 13
- 14. 24 Node HDD 25 Gbe Network Avg Read Throughput (GET) Avg Write Throughput (PUT) Distributed 16.3 GB/s 9.4 GB/s S3 Benchmark - Spinning Disk: 24 Nodes 14
- 15. 8 Node NVMe 100 Gbe Network Avg Read Throughput (GET) Avg Write Throughput (PUT) Distributed 46.54 GB/s 34.4 GB/s Distributed with Encryption 46.4 GB/s 34.6 GB/s S3 Benchmark - NVMe: 8 Nodes 15
- 16. 32 Node NVMe 100 Gbe Network Read Throughput (GET) Write Throughput (PUT) Distributed 183.2 GB/s 171.3 GB/s Distributed with Encryption 162 GB/s 114.7 GB/s S3 Benchmark - NVMe: 32 Nodes 16 The addition of a separate network for internode traffic would effectively double the performance.
- 17. Linear Scalability HDD Scalability NVMe Scalability 8.57 10.81 9.4 17 20 15 10 5 0 16 node 24 node Maximum Bandwidth (GB/sec) PUT GET Trendline 200 150 100 50 0 8 node 32 node Maximum Bandwidth (GB/sec) 34.4 46.5 171.3 183.2 PUT GET Trendline
- 19. 19 Spark and Presto: MinIO vs. AWS
- 20. Why MinIO is so Fast SINGLE LAYER We are a single layer, object only. Multiple layers cause latency, complexity. SIMD ACCELERATION By writing the core parts of MinIO in assembly language (SIMD extensions) we are hyperfast on commodity HW. COMBINATION OF GO + GOASM Delivering C-like performance by combining GO + Assembly Language and targeting them to the task. NO METADATA DATABASE By writing object and metadata together you make all operations single and atomic. Multiple steps for other vendors.
- 21. 21 Big Data/Machine Learning environments HDFS replacements High performance data lake/warehouse infrastructure Cloud native applications (replacing file + block) Multi-cloud environments (portability) Endpoint for streaming workloads By deploying for performance, long term archival storage and disaster recovery are effectively free. MinIO Deployment Use Cases
- 22. 22 Takeaways MinIO has grown into the dominant provider of object storage in the private cloud. Performance and scalability are key criteria and MinIO is leading the market on both fronts. We are entering the Kubernetes-era of computing and no object storage solution is is better positioned than MinIO.
- 23. 23
- 25. P P P P SERVER 2 SERVER 3 SERVER 32SERVER 1 OBJECT 1 OBJECT 2 100 GBe SWITCH S3 API minio server http://host{1...32}/export{1...32} MinIO Architecture
- 26. 26 At the Center of a Modern Architecture S3 SELECT SQL S3 S3 SELECT SQL S3 SELECT Data Processing Applications Streaming Data Events Logs Sensor Data Social Transaction Data Infrastructure Machine Learning STATELESS
- 27. 27 MinIO/Kubernetes Deployment Scenarios Stateful SetsDeployment M-Operator M3 (mcube)
- 28. 28 VMware’s Standard for K8’s Storage
- 29. Kubernetes Storage Leadership Source: CNCF Survey 2018 Kubernetes related Dockerhub image downloads: Source: Torsten Volk, EMA Which of these cloud native storage projects is your company / organization using:
- 30. 30 MinIO on NAS & Cloud (gateway mode) S3 NAS / Cloud Server NFSNFS APPLICATIONS
- 31. 31 Splunk SmartStore and MinIO
- 32. 32 Performance Oriented, Enterprise Grade Features Amazon S3 API is the de facto standard for object storage. MinIO implements Amazon S3 v2/v4 API. MinIO has a pluggable back-end with integrations with external storage backends such as NAS, Google Cloud Storage, and Azure Blob Storage. MinIO server triggers Lambda functions through event notification service. OCR, audit compliance are good examples of lambda computing. You may lose up to half the number of drives and still recover from it. Data protection code is accelerated using SIMD instructions on x64 and ARM CPUs. Federation allows combining unlimited number of MinIO instances to form a single global. Disk-cache feature allows content to be closer to the tenants. This enables objects to be delivered with high performance and reduced bandwidth cost. MinIO Server integrates with Identity Providers such as WSO2, Keycloak, Okta, Ping Identity to allow applications or users to authenticate and use Object Storage. MinIO provides confidentiality, integrity and authenticity assurances for encrypted data with negligible performance overhead. MinIO supports WORM (Write-Once-Read-Many) for long-term data retention as mandated by regulations or compliance rules. S3 Select filters the contents of an object using SQL to parse object data into records, and returns only records that match the query.
- 33. Scalability DATA CENTER 1 DATA CENTER 2 DATA CENTER 3 ETCD US-EAST US-WEST DNS EU-1 EU-2 ▪ ▪ ▪
- 34. Continuous Disaster Recovery mc mirror --watch --recursive dc1/ dc2/ PRIMARY SITE SECONDARY SITE network
- 35. OBJECTS ELASTICSEARCH POSTGRESQL REDIS AMQP NATS IO WEBHOOKS MQTT LAMBDA FUNCTION BATCH PROCESS AUDIT LOGS THUMBNAIL OCR COMPRESS MinIO Lambda Functions
- 36. Identity & Access Management Client grant OpenID Connect Token Get object Token STS Temporary credentials IDENTITY PROVIDER (IdP) APPLICATION
- 37. MinIO Encryption SSE-S3 SSE-C My Bucket My Object Data KMS Random IV Algorithm Name Sealed Object Key Random IV Sealed Object Key KMS Key ID Sealed KMS Data Key Algorithm Name Metadata Metadata
- 38. Disk1 MyBucket MyObject part.1 export-xl/ xl.json { “version”: “1.0.1”, “format”: “xl”, “stat”: { “size”: 2286, “modTime”: “2017-12-02T00:24:20.975968336Z”, }, “erasure”: { “algorithm”: “klauspost/reedsolomon/vandermonde”, “data”: 2, “parity”: 2, “blockSize”: 10485760, “index”: 2, “distribution”: “[ 2, 3, 4, 1 ], “checksum”: [ { “name”: “part.1”, “algorithm”: “blake2b”, “hash”: “c24fa0451fd85a3a482c...b672b7f” } ] }, “minio”: { “release”: “DEVELOPMENT.GOGET” }, “meta”: { “content-type”: “application/octet-stream”, “etag”: “c1d217c52d44c9eab00e81496b2b91b6” }, “parts”: [ { “number”: 1, “name”: “part.1”, “etag”: “”, “size”: 2286 } ] } Disk2 MyBucket MyObject part.1 xl.json Disk3 MyBucket MyObject part.1 xl.json Disk4 MyBucket MyObject part.1 xl.json Erasure Code Internals
- 40. MinIO Encryption - SSE-C MyBucket S3 Client HTTPS HTTP Body HTTP Headers X--Amz-...-Customer-Algorithm X-Amz-...-Customer-Key X-Amz-...-Customer-Key-MD5 My Object Generated randomly Object Key SSE Key Sent by client Generated randomly IV Sealed Object Key Object Name Bucket Name IV Algorithm Name Sealed Object Key Metadata Object Data
- 41. MinIO Encryption - SSE-S3 My Bucket S3 Client HTTPS HTTP Body HTTP Headers X-Amz-Server-Side- Encryption: AES256 My Object Generated randomly Generated randomly IV Algorithm Name Sealed Object Key Metadata Object Data KMS KMS Sealed Key KMS Key ID Object Key KMS Data Key IV Sealed Object Key Object Name Bucket Name KMS Sealed Key Master Key 1