DockerCon 18 Cool Hacks: solo.io
7 likes1,138 views
This document discusses the challenges of transitioning between monolithic, microservices, and serverless architectures. It presents several potential solutions to this problem, ultimately advocating for a hybrid approach that allows applications to utilize elements of each architecture. The key benefits of this hybrid app vision are fast delivery of new features, a gradual transition with minimal disruption, and a unified toolset. The document demonstrates this approach using Gloo, an open source project that acts as a layer to connect applications built with different architectures.
DockerCon 18 Cool Hacks: solo.io
- 1. MONOLITHIC TO MICROSERVICES TO SERVERLESS
- 2. THE PROBLEM: Disparate ecosystems, hard transition Monolithic Apps Microservices Serverless Ansible Docker EE AWS ! APM Splunk SOA Prometheus OpenTracing microservices CloudWatch X-RAY Event-driven Enterprise faces 4 main problems in adopting innovative architectures: 1. Insolation between brown and green field 2. Transition is lengthy and diverts essential personnel from core mission 3. Duplicate redundant tools 4. Requires silo teams
- 3. POTENTIAL SOLUTION I: Stay monolithic Monolithic Apps Microservices Serverless Pros: minimum perturbation. Cons: slow discontinuous delivery of new features, slow and costly maintenance.
- 4. POTENTIAL SOLUTION II: Keep your monolithic, use microservices/ serverless only for new applications Monolithic Apps Microservices Serverless Pros: new applications benefit from the advantages of modern architectures Cons: core business still suffers from slow delivery of new features and costly maintenance. Silo teams, redundant tooling. Existing Apps New Apps
- 5. Monolithic Apps Microservices Serverless Pros: if and when complete, all the advantages of cloud-native architectures. Cons: takes away key personnel from critical missions, no new features for a while; risky: low probability of success. POTENTIAL SOLUTION III: ”Big-bang” refractoring the monolithic application
- 6. OUR VISION: Hybrid App Monolithic Apps Microservices Serverless Ansible Docker EE AWS ! APM Splunk SOA Prometheus OpenTracing microservices CloudWatch X-RAY Event-driven
- 7. OUR VISION: Hybrid App Monolithic Apps Microservices Serverless Ansible Docker EE AWS ! Access all available functions, the smallest unit of compute in all architectures
- 8. OUR VISION: Hybrid App Monolithic Apps Microservices Serverless Ansible Docker EE AWS ! Access all available functions, the smallest unit of compute in all architectures Provide a unified ecosystem, allowing the IT team • To seamlessly combine existing functionalities • Deploy, schedule, monitor, debug, and log applications using a single toolbox Hybrid app
- 9. OUR VISION: Hybrid App Monolithic Apps Microservices Serverless Ansible Docker EE AWS ! Pros: fast delivery of new features in modern architectures; stop “digging the hole”; gradual transition = minimal interruption; unified tooling. Hybrid app
- 10. DEMO
- 11. Gloo will glue together your hybrid environment Event Request Emit Docker Enterprise Edition Platform Monolithic Apps Microservices Portable Serverless Proprietary Serverless Function-level routing: canary deployment, security, caching Data plane: builds around and extends the Envoy proxy (C++) Control plane: pluggable, extensible architecture (GO) Centralized place to manage security, observability, trafficEnvoy Proxy
- 12. Gloo will glue together your hybrid environment Event Request Emit Docker Enterprise Edition Platform Monolithic Apps Microservices Portable Serverless Proprietary Serverless Full integration with all components of your environment: all clouds, all platforms, all technologies Automatic discovery of all your functions, services & appsEnvoy Proxy
- 13. Let Gloo build your hybrid apps Docker Enterprise Edition Platform Monolithic Apps Microservices Portable Serverless Proprietary Serverless Hybrid app + - ÷× ÷ -x + calc.com/add calc.com/sub calc.com/divcalc.com/mul Envoy Proxy
- 14. Squash Demo How do we debug a super-distributed app?
- 15. One more thing…
- 17. GraphQL Server GraphQL Arbitrary Code Client Arbitrary Code Arbitrary Code Resolvers Validation Caching Logging Metrics Security Aggregation To build GraphQL server:
- 18. GraphQL with Gloo Resolvers Validation Caching Logging Metrics Security Aggregation GraphQL Arbitrary Code Client Arbitrary Code Arbitrary Code To build GraphQL server: GraphQL Client Function routing Validation Caching Logging Metrics Function-level security Aggregation Let Gloo do the heavy- lifting
- 19. GraphQL Client The QLoo Server: üLoggings üMetrics üCaching üFunctions Routing üValidation üFunction level Security üFunction discovery üAggregation INTRODUCING QLOO: The Codeless GraphQL Server
- 20. Flow With QLoo query avg($x: Float, $y:Float) { sum (input1: $x, input2: $y) { divide_by_2 { result } } } avg sum div_by_2 avg On prem Google cloud AWS + x ÷ /2 -
- 21. DEMO
- 22. CODELESS GRAPHQL SERVER NOW OPEN: HTTPS://GITHUB.COM/SOLO-IO/QLOO
- 23. We Are Hiring! HAPPY ING &ING, ING! GitHub: https://github.com/solo-io Twitter: @getsoloio, @idit_levine Slack: https://slack.solo.io