Containerized Storage for Containers
0 likes475 views
The document discusses a presentation on containerized storage solutions by Murat Karslioglu at the Bay Area Kubernetes Meetup, focusing on OpenEBS and its benefits for managing storage in Kubernetes environments. It highlights the advantages of Direct Attached Storage (DAS) over traditional storage methods, addressing challenges in storage management and performance while advocating for a dynamic and efficient approach to cloud infrastructure. Key takeaways include the need for micro-service granularity, the importance of resource efficiency, and the potential for cross-cloud portability.
Containerized Storage for Containers
- 1. Bay Area Kubernetes Meetup – December 14 Murat Karslioglu VP of Solutions @OpenEBS @muratkarslioglu Blog: www.containerized.me Containerized Storage for Containers
- 2. Who? v Team of storage and DevOps engineers with a 5 year track record of delivering containerized storage (built on jails) to NTT, eSilicon, TCS and other service providers. MayaData sponsors OpenEBS. 2017 MayaData Inc. All Rights Reserved
- 3. 2017 Storage Developer Conference. © CloudByte Inc.. All Rights Reserved. 3 2017 CloudByte Inc. All Rights Reserved What if the drivers disrupting the storage industry helped to fix storage – and data management – once for all? DevOps Containerization Docker AWS Microservices OpenSource Non-volatile Memory (NVMe) Fear of lock-in
- 4. 2017 Storage Developer Conference. © CloudByte Inc.. All Rights Reserved. 4 2017 CloudByte Inc. All Rights Reserved Cross cloud infrastructure becomes a reality – and data management is fixed once and for all. And storage concerns just fade away for containerized enterprises.
- 5. 2017 Storage Developer Conference. © CloudByte Inc.. All Rights Reserved. 5 2017 CloudByte Inc. All Rights Reserved What if storage for container native applications was itself container native?
- 6. DAS increasingly preferred to NAS/SAN DAS Benefits: Simple Ties application to storage Predictable for capacity planning App deals with resiliency Can be faster Concerns: Under-utilized hardware ○ 10% or less utilization Wastes data center Difficult to manage Lacks storage features Cannot be repurposed - made for one workload Does not support mobility of workloads via containers Cross cloud impossible Other example DAS users moving to Kubernetes: 2017 MayaData Inc. All Rights Reserved “We have ~100k nodes of Cassandra alone, and use DAS because it is easier even if it burns energy and capEx.”
- 7. DAS increasingly preferred to NAS/SAN DAS Benefits: Simple Ties application to storage Predictable for capacity planning App deals with resiliency Can be faster Concerns: Under-utilized hardware ○ 10% or less utilization Wastes data center Difficult to manage Lacks storage features Cannot be repurposed - made for one workload Does not support mobility of workloads via containers Cross cloud impossible Other example DAS users moving to Kubernetes: 2017 MayaData Inc. All Rights Reserved “We have ~100k nodes of Cassandra alone, and use DAS because it is easier even if it burns energy and capEx.”
- 8. DAS and Distributed DAS Benefits: Simple Ties application to storage Predictable for capacity planning App deals with resiliency Can be faster Concerns: Under-utilized hardware ○ 10% or less utilization Wastes data center Difficult to manage Lacks storage features Cannot be repurposed - made for one workload Does not support mobility of workloads via containers Cross cloud impossible Distributed Benefits: Centralized management Greater density and efficiency Storage features such as: ○ Data protection ○ Snapshots for versioning Concerns: Additional complexity Enormous blast radius Expensive Requires storage engineering Challenged by container dynamism No per microservice storage policy I/O blender impairs performance Locks customers into vendor Cross cloud impossible 2017 MayaData Inc. All Rights Reserved
- 9. DAS and Distributed DAS Benefits: Simple Ties application to storage Predictable for capacity planning App deals with resiliency Can be faster Concerns: Under-utilized hardware ○ 10% or less utilization Wastes data center Difficult to manage Lacks storage features Cannot be repurposed - made for one workload Does not support mobility of workloads via containers Cross cloud impossible Distributed Benefits: Centralized management Greater density and efficiency Storage features such as: ○ Data protection ○ Snapshots for versioning Concerns: Additional complexity Enormous blast radius Expensive Requires storage engineering Challenged by container dynamism No per microservice storage policy I/O blender impairs performance Locks customers into vendor Cross cloud impossible 2017 MayaData Inc. All Rights Reserved
- 10. DAS Under-utilized hardware Hard to manage Lacks all storage features Hard to support workload mobility 2017 MayaData Inc. All Rights Reserved
- 11. YASSS: yet another scale out storage system 2017 MayaData Inc. All Rights Reserved
- 12. YASSS: yet another scale out storage system Huge blast radius 2017 MayaData Inc. All Rights Reserved
- 13. !YASSS Yet Another Scale Out Storage System 2017 MayaData Inc. All Rights Reserved
- 14. !YASSS Yet Another Scale Out Storage System 2017 MayaData Inc. All Rights Reserved
- 15. Container Attached Storage = DAS++ DAS Benefits: Simple Ties application to storage Predictable for capacity planning App deals with resiliency Can be faster Concerns: Under-utilized hardware ○ 10% or less utilization Wastes data center Difficult to manage Lacks storage features Cannot be repurposed - made for one workload Does not support mobility of workloads via containers Cross cloud impossible OpenEBS = “CAS” ✓ Simple ✓ No new skills required ✓ Per microservice storage policy ✓ Data protection & snapshots ✓ Reduces cloud vendor lock-in ✓ Eliminates storage vendor lock-in ✓ Highest possible efficiency ✓ Large & growing OSS community ✓ Natively cross cloud Distributed Benefits: Centralized management Greater density and efficiency Storage features such as: ○ Data protection ○ Snapshots for versioning Concerns: Additional complexity Enormous blast radius Expensive Requires storage engineering Challenged by container dynamism No per microservice storage policy I/O blender impairs performance Locks customers into vendor Cross cloud impossible 2017 MayaData Inc. All Rights Reserved CAS DAS Distributed
- 16. Container Attached Storage No special skills needed! 2017 MayaData Inc. All Rights Reserved
- 17. OpenEBS metrics 2017 MayaData Inc. All Rights Reserved
- 18. Converting stars into users 2017 MayaData Inc. All Rights Reserved
- 19. Let’s have fun! 2017 MayaData Inc. All Rights Reserved
- 20. Let’s have fun! 2017 MayaData Inc. All Rights Reserved
- 21. Architecture: Kubernetes K8S Master Minion POD Container Container Container Kubele t POD Container Container Container Kubele t POD Container Container Container Kubele t Minion POD Container Container Container Kubele t POD Container Container Container Kubele t POD Container Container Container Kubele t Minion POD Container Container Container Kubele t POD Container Container Container Kubele t POD Container Container Container Kubele t etcd APIs Cntrl Schl d Minions run on physical nodes PODs group containers, share an IP address, and each include a Kubelet agent K8S Master services include: etcd, APIs, the scheduler, the control manager & others 2017 MayaData Inc. All Rights Reserved
- 22. Architecture: Kubernetes + OpenEBS K8S Master Minion POD Container Container Container Kubele t POD Container Container Container Kubele t POD Container Container Container Kubele t Minion POD Container Container Container Kubele t POD Container Container Container Kubele t POD Container Container Container Kubele t Minion POD Container Container Container Kubele t POD Container Container Container Kubele t POD Container Container Container Kubele t etcd APIs Cntrl Schld OpenEBSAPIs Schld Data Containers run in PODs on physical machines Data Containers group physical disks local or remote, manage replicas, provide QoS control & more OpenEBS brain runs on the Master; delivers services such as: APIs, a storage scheduler, analytics… 2017 MayaData Inc. All Rights Reserved
- 23. Architecture: Kubernetes + OpenEBS + Maya K8S Master Minion POD Container Container Container Kubele t POD Container Container Container Kubele t POD Container Container Container Kubele t Minion POD Container Container Container Kubele t POD Container Container Container Kubele t POD Container Container Container Kubele t Minion POD Container Container Container Kubele t POD Container Container Container Kubele t POD Container Container Container Kubele t etcd APIs Cntrl Schld OpenEBSAPIs Schld Data Containers run in PODs on physical machines Data Containers group physical disks local or remote, manage replicas, provide QoS control & more OpenEBS brain runs on the Master; delivers services such as: APIs, a storage scheduler, analytics… Maya delivered as Cloud Converged Appliance & SaaS solution to fully automate storage operations 2017 MayaData Inc. All Rights Reserved
- 24. Why per micro-service granularity important? Today’s scale out storage systems are almost infinitely configurable. Matching their configurations to static workloads is difficult. In a world of dynamic, ever evolving workloads due to micro-services and multi-cloud deployments, it is not a tractable problem for humans. Let’s say you want to benchmark for your workload: Source: Alex Aizman’s blog: https://storagetarget.com/author/alexaizman/ 2017 MayaData Inc. All Rights Reserved
- 25. Today’s scale out storage systems are almost infinitely configurable. Matching their configurations to static workloads is difficult. In a world of dynamic, ever evolving workloads due to micro services and multi-cloud deployments, it is not a tractable problem for humans. Let’s say you want to benchmark for your workload: Duration of run 6,700 years Source: Alex Aizman’s blog: https://storaget arget.com/auth or/alexaizman/ Why per micro-service granularity important? 2017 MayaData Inc. All Rights Reserved
- 26. Storage Orchestration - K8s Vs OpenEBS Functionality K8s OpenEBS Volume Life-cycle - provisioning, attaching/detach, de-provisioning ✓ Monitoring the volume data ✓ Scheduling the volume provisioning intelligently ✓ Smart space management ✓ Availability of volume ✓ Application aware backup/snapshots ✓ QoS guarantee (latency guarantee) ✓ Smart movement of the volumes across racks and regions ✓ Cataloging of storage requirements for different applications ✓ 2017 MayaData Inc. All Rights Reserved
- 27. Normal behavior POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet Data Center 1 Maya POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet Data Center 2 POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet EBS AWS EBS EBS EBS POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet GPD GCE GPD GPD GPD 1. CI/CD pipeline deploys: 2. OpenEBS reads intent FROM K8S For example: - 2 sync copies - 50 / 50 cloud - 2 Data centers 2017 MayaData Inc. All Rights Reserved
- 28. Normal behavior POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet Data Center 1 Maya POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet Data Center 2 POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet EBS AWS EBS EBS EBS 1. CI/CD pipeline deploys: 2. OpenEBS reads intent FROM K8S For example: - 2 sync copies - 50 / 50 cloud - 2 Data centers POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet GPD GCE GPD GPD GPD 2017 MayaData Inc. All Rights Reserved
- 29. Normal behavior POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet Data Center 1 Maya POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet Data Center 2 POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet EBS AWS EBS EBS EBS 1. CI/CD pipeline deploys: 2. OpenEBS reads intent FROM K8S For example: - 2 sync copies - 50 / 50 cloud - 2 Data centers 3. writes STORED locally: - working set in NVM - rebalancing in background POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet GPD GCE GPD GPD GPD 2017 MayaData Inc. All Rights Reserved
- 30. 2017 MayaData Inc. All Rights Reserved Normal behavior: differentiation POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet Data Center 1 Maya POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet Data Center 2 POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet EBS AWS EBS EBS EBS 1. Transparency: - works across clouds, & CAS, seamlessly 2. FAST: - Much faster than DAS or network attached - Could be 100 or even 1000x faster POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet GPD GCE GPD GPD GPD
- 31. 2017 CloudByte Inc. All Rights Reserved Normal behavior: differentiation POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet Data Center 1 Maya POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet Data Center 2 POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet EBS AWS EBS EBS EBS 1. Transparency: - works across clouds, & CAS, seamlessly 2. FAST: - Much faster than DAS or network attached - Could be 100 or even 1000x faster 3. Happier Cassandra: - When Cassandra nodes fail, rebalancing, which degrades performance, typically NOT needed POD Container Container ContainerKubelet POD Container Container ContainerKubelet POD Container Container ContainerKubelet GPD GCE GPD GPD GPD 4. Happier CFO: - Far greater utilization - No more cloud lock-in
- 32. Maya: Cross-cloud control plane ○ Visibility, automation, collaboration, and, over time, learning via machine learning ○ OpenEBS users can subscribe to a free version and then are upsold to a subscription that includes OpenEBS enterprise support OpenEBS: Containerized Storage for Containers ○ Open source software that allows each workload - and DevOps team - to have their own storage controller
- 33. Personas 2017 MayaData Inc. All Rights Reserved
- 34. API MAYAOnline.io ✓ Visibility ✓ ChatOps ✓ Optimization ChatOps MAYA GUI
- 35. 2017 CloudByte Inc. All Rights Reserved DEMO TIME
- 36. 2017 CloudByte Inc. All Rights Reserved Install OpenEBS
- 37. 2017 CloudByte Inc. All Rights Reserved Install OpenEBS kubectl apply -f https://openebs.github.io/charts/openebs-operator.yaml
- 38. 2017 CloudByte Inc. All Rights Reserved Install PostgreSQL Cluster on OpenEBS volumes
- 39. 2017 CloudByte Inc. All Rights Reserved Install PostgreSQL Cluster on OpenEBS volumes git clone https://github.com/openebs/openebs.git cd openebs/k8s/demo/crunchy-postgres ./run.sh
- 40. Key take-aways OpenEBS always available multi-cloud block services are delivered via containerized microservices Granularity: Each pod has its own block protocol stack, tiering engine, QoS engine and more. SLAs per POD. Performance & efficiency: access patterns maintained per-volume. Hot-data on NVMe-Flash or on 3DX-Memory. Cold- data on slower disks / SAN/ Cloud-Storage/S3. Scalability: Metadata based on “size of the volume” & not on “number of volumes”. Can reduce metadata traversal 100x. Blast radius: Eliminates risk of scale out storage system corruption. And much, much more such as cost savings due to resource efficiency, innate multitenancy, and cross cloud portability The Conway’s Law benefits might be the most important - innovation unlocked ✓ ✓ ✓ ✓ 2017 MayaData Inc. All Rights Reserved
- 41. Q&A Now start hacking! Find out HOW it works! www.openebs.io / www.mayaonline.io Slack.openebs.io blog.openebs.io Github.com/openEBS Follow @openebs Me: @muratkarslioglu 2017 MayaData Inc. All Rights Reserved