The Fastest Way to Redis on Pivotal Cloud Foundry
1 like4,519 views
The document discusses Redis as a NoSQL solution for cloud-native developers in the context of Pivotal Cloud Foundry. It highlights the advantages of Redis, including its performance as an in-memory key-value store and the benefits of automating its deployment as a service to enhance developers' productivity. Additionally, it outlines features of Redis for Pivotal Cloud Foundry 1.8, including dedicated and shared VM instances, customizable service plans, and operator-managed backups.
![Maximize developer choice while providing significant automation
Operator & Developer Lifecycle
Operator DeveloperRedis for Pivotal
Cloud Foundry
Enable and configure
on-demand plans
Enable and configure
pre-provisioned plans
Install Redis Create service broker VM
Broker registers plans as
available in marketplace
[Dedicated-VM/Shared-VM only]
Create configured number of
dedicated VM instances
Run smoke tests](https://image.slidesharecdn.com/fastestwaytoredisonpivotalcloudfoundry-170629005258/85/The-Fastest-Way-to-Redis-on-Pivotal-Cloud-Foundry-18-320.jpg)
![$ CF create-service [service] [plan] [instance name]
Create or allocate Redis instance
if quota or pre-configured
instances are available.
App gains access to Redis
service instance
$ CF bind-service [app name] [instance name]
App access to Redis service
instance is revoked
$ CF unbind-service [app name] [instance name]
Redis service resource is
destroyed or pre-provisioned
instance is flushed and pooled.
$ CF delete-service [instance name]
Service broker ready
Update service Update service broker VM
Update all service instances
Run smoke tests
Service broker ready](https://image.slidesharecdn.com/fastestwaytoredisonpivotalcloudfoundry-170629005258/85/The-Fastest-Way-to-Redis-on-Pivotal-Cloud-Foundry-19-320.jpg)
The Fastest Way to Redis on Pivotal Cloud Foundry
- 1. The Fastest Way to Redis in Pivotal Cloud Foundry Greg Chase, Product Marketing Craig Olrich, Platform Architect
- 2. Agenda 1. Redis, NoSQL and Cloud-Native Developers 2. Platform for Delivery of Redis as a Service 3. Design of Redis for PCF 4. Demo and roadmap of Redis for Pivotal Cloud Foundry 5. Q&A
- 3. Redis is a Unique NoSQL Offering for Cloud-Native Developers
- 4. Open source key-value store used as a cache, data store, and queue Redis is an In-Memory NoSQL Data Structure Server ● In-memory: entire data set is held in RAM ● Persistence: journal of changes to in-memory data set is persisted to disk for durability ● Data types as keys: ○ Strings ○ Lists ○ Sets ○ Sorted sets ○ Hash tables ○ HyperLogLogs ○ GeoHash ● 49 Supported languages (including): C, C++, C#, Go, Haskell, Java, Javascript, Objective-C, Perl, PHP, Python, R, Ruby, Rust, Scala
- 5. Third most popular NoSQL engine, and #1 in-memory key-value store Redis is Important For Your Developers’ Toolbox ● Has very different data model from other popular engines ● Much more common than other in-memory KV stores ● Great for caching & low latency data management
- 6. Delivering Redis as a Service Requires a Cloud-Native Platform
- 7. 7 Software Cycles Need to Get Faster Teams are trying to release faster with continuous delivery and devops Sources: 1) How Netflix Deploys Code, InfoQ 2) Continuous Delivery + DevOps, Nicole Forsgren, Agile India, 2016 Service Mindset Time Risk Install Mindset Time Risk Cloud-Native Practices Infrequent (Manual) Delivery Measured in months Continuous (automated) Delivery Measured in pushes per day Don’t let manual provisioning of systems get in the way of agile delivery at scale
- 8. 8 How DIY Redis Impacts Software Cycles One datastore for many apps Provisioning systems Build and configure own data stack Inconsistent updates Inconsistent deployment testing Inconsistent backups Inconsistent security settings Slow vulnerability patching Build & Deliver Production Snowflake datastores case gaps in continuous delivery at all stages Picking right system sizing
- 9. Speed up software delivery by spending less time on overhead activities Help Your Developers Focus on Value Source: “Electric Cloud LinkedIn Survey to software developers” Average Hours/Week Business Value Overhead 0 3.25 6.5 9.75 9.75 Average hours spent per week by developers on tasks
- 10. Build apps faster with lots of choice; deliver faster with pervasive automation Pivotal Cloud Foundry Allows Focus on Value Languages Clouds Platform Services OSes MySQL Dynatrace AppMon Spring Cloud Services Single Sign On ✓ Consistent Contracts ✓ Fully Automated, Repeatable platform managed DevOps processes ✓ Developer + Ops Friendly Constructs ✓ Easy to move/re-architect apps ✓ Infrastructure Failure Agnostic Pervasive Automation Automated, Fast, Flexible, Safe Developer Choice
- 11. 11 cf push ~45 seconds ~15+ Days Find available hosts Install & configure runtime Install & configure middleware Pull application source code Retrieve dependent libraries Create application package Install, configure dependent service(s) Deploy container to host(s) Load environment variables Configure load balancer Configure firewalls Update service monitoring tools Configure log collector 2 Days 1 Day 1 Day ¼ Day ¼ Day ¼ Day 2 Days ½ Day ¼ Day 2 Days 2 Days 3 Days 1 Day Code Complete & Tested Application in Production Speed & Consistency Example Automation of App Deployment
- 12. Delivering Enterprise-Ready Redis in a Cloud-Native Platform
- 13. Services leverage higher degrees of automation depending on APIs implemented Levels of Service Integration in Pivotal Cloud Foundry Unmanaged Services Managed Services On Demand Services ● External resources ● Automates application binding ● Great for remote services ● Pre-provisioned BOSH- managed resources ● Automates application binding ● Great for shared services ● Dynamically provisioned by developers ● Automates service creation and application binding ● Used for dedicated instances Example Services:
- 14. Maximize developer choice while providing significant automation Design Philosophy of Pivotal Data Services Dedicated & On Demand Enterprise Ready Use-case tailored On-demand, self-service provisioning of isolated instances in any cloud. Operator defined plans, upgrade rules, and quotas that meet enterprise requirements Convenient packaging for developers
- 15. Now providing on-demand, dedicated instances of Redis New with Redis for Pivotal Cloud Foundry 1.8 Developers Operators Sm plan Med plan Lrg planBApp Binding ● Operators create plans with varied VM and disk sizes to allow App Developers to choose the plan that fits their needs ● Plans are configured to enable caching with persistence and have operator-set quotas to manage resource use ● App Developer self-service to provision instances, removing transactional overhead from the Operations team ● Operators and App Developers can tailor certain Redis configurations to ensure instances fit use cases
- 16. Developers can access pre-provisioned single VM instances and shared VM instances. Redis for Pivotal Cloud Foundry 1.8 – Additional Developers Operators App Binding B ● Operators pre-provision a certain number of set VM and disk sizes for app developer usage ● Plans support in-memory data store scenario with persistence and automated backups with restore capabilities ● App developers can allocate pre-provisioned instances via self-service ● Shared-VM plan offers several instances within a VM to minimize resource use during development App Binding Pre-provisioned, dedicated VMs Pre-provisioned, shared VM B B B
- 17. Enterprise-ready Redis as a Service Redis for Pivotal Cloud Foundry 1.8 Features Provisions Redis 3.2.8 Platform Integration Features ✓ Dynamically created, isolated Redis instances ✓ Pre-provisioned dedicated VM instances of Redis ✓ Pre-provisioned shared VM for small Redis instances ✓ Automated updates and upgrades by operator ✓ Customizable service plans ✓ Operator-set plan quotas ✓ Developers can migrate between plans ✓ “Secure by default” deployment ✓ Works with IPsec plugin for encrypted communication ✓ Deployment smoke tests ✓ Metrics and monitoring with Loggregator and syslog ✓ Operator managed backup policies and database restores
- 18. Maximize developer choice while providing significant automation Operator & Developer Lifecycle Operator DeveloperRedis for Pivotal Cloud Foundry Enable and configure on-demand plans Enable and configure pre-provisioned plans Install Redis Create service broker VM Broker registers plans as available in marketplace [Dedicated-VM/Shared-VM only] Create configured number of dedicated VM instances Run smoke tests
- 19. $ CF create-service [service] [plan] [instance name] Create or allocate Redis instance if quota or pre-configured instances are available. App gains access to Redis service instance $ CF bind-service [app name] [instance name] App access to Redis service instance is revoked $ CF unbind-service [app name] [instance name] Redis service resource is destroyed or pre-provisioned instance is flushed and pooled. $ CF delete-service [instance name] Service broker ready Update service Update service broker VM Update all service instances Run smoke tests Service broker ready
- 20. Provide Redis as a service for your developers in any cloud Use-case Tailored Redis Plans Net new Redis Apps Replatformed Redis Apps Caching In-memory data store Backing Service for Spring Services & projects
- 21. How On Demand Redis Speeds Software Cycles Automated provisioning of dedicated instances aids continuous delivery Example Concourse pipeline utilizing Redis for Pivotal Cloud Foundry self-service provisioning for automated testing:
- 22. Focus on value developing data-driven applications Benefits of Redis for Pivotal Cloud Foundry Benefits for Developers ➢ Focus time on coding, not data storage administration ➢ Get fast access to systems ➢ Isolated from noisy neighbors ➢ Flexible control when needed ➢ Choice of system configurations suited for use cases ➢ Visibility into data store events & operations ➢ Data is safely backed up Benefits for Operators ➢ Focus time on improving platform, not patching and administration ➢ Scale operations to support more apps and developers ➢ Minimal time to update and patch all deployed databases ➢ Bugs and vulnerabilities are patched quickly ➢ Efficient consumption of resources ➢ Consistently enforced security ➢ Avoid data loss & recover quickly
- 23. Demo 2
- 24. Demo Redis App User Developer Push the app Create Redis service instance Bind service instance to app App is ready View app interface App in PCF Redis service instanceCheck for values in cache Returned cache value, if any Cached values expire over time Market data API No cached value? Fetch new data Update cache with fetched values
- 25. Quarterly release cadence Futures for Redis for Pivotal Cloud Foundry First on-demand version Q1 Q2 Q3 Q4 1.8 CY 1.10 1.11 1.12 1.13 2018+ On-demand datastore plan Migration from pre-configured to on-demand Further enhancement of on-demand instances 1.9 1.10 1.11
- 26. Note: Redis for Pivotal Cloud Foundry 1.9 is due to be released at the end of the month Where to get more information ● Read the documentation ● Download Redis for Pivotal Cloud Foundry on Pivotal Network ● Try out Redis in a Pivotal Cloud Foundry Developer’s Sandbox
- 27. Any questions? 2
- 28. Demand that your datastores be “on demand”! Redis for Pivotal Cloud Foundry