Well Architected Java Backends

The innovations in cloud computing brings many new options for Java architects. Architecture styles range from traditional monoliths to container-based microservices or even "serverless" functions. But the possibilities of services and stacks are so many that choosing the most adequate one can be overwhelming. This presentation is a guide through common implementation patterns for Java applications in the cloud.

The AWS Well Architected Framework will be used as a reference assessment of security, reliability, peformance efficiency, cost optimization and operational excellence. Although each organization will have its own peculiar needs, all share at least some concern for these key requirements. This talk presents reference implementations, tools and samples, focused with the peculiarities of the Java platform and frameworks.

Join this practical "state of the union" presentation and let's share practices and examples for well architected java backends.

Transcript

1. Julio Faerman @julioaws Well Architected Java Backends
2. How do you demonstrate your architecture is ready?
3. “… we are using [popular stack] like [successful company]…”
4. “… we are an enterprise compliant with [standard audit]…”
5. “… and we have automated testing…”
6. “… but is that enough?” ¯\_(
   )_/¯ #FOMO
7. Optimizing Subjective Qualitative Ephemeral Private Satisficing Objective Quantitative Durable Shared
8. Optimizing Satisficing latency P99 < 250ms throughput 42 tps
9. Subjective Objective HTTPS Everywhere AES-GCM with Key Derivation Perfect Forward Secrecy Multi-Factor Authentication Temporary Credentials Live Log Analytics WAF Automation DDoS Protection Dependency Inspection CVE Agent …
10. Qualitative Quantitative High Availability month_uptime > 99.900% = 99.915%
11. Ephemeral <dependency> <groupId>org.pegdown</groupId> <artifactId>pegdown</artifactId> <version>1.6.0</version> <scope>test</scope> </dependency>
12. Durable
13. Shared
14. Operational Excellence Security Reliability Performance Efficiency Cost Optimization AWS Well Architected Framework Pillars
15. Operational Excellence Security Reliability Performance Efficiency Cost Optimization AWS Well Architected Framework SEC 1: How are you protecting access to and use of the AWS account root user credentials?
16. Operational Excellence Security Reliability Performance Efficiency Cost Optimization AWS Well Architected Framework REL 8: How are you testing your resiliency?
17. Operational Excellence Security Reliability Performance Efficiency Cost Optimization AWS Well Architected Framework PERF 2: How did you select your compute solution?
18. Operational Excellence Security Reliability Performance Efficiency Cost Optimization AWS Well Architected Framework COST 4: How do you make sure your capacity matches but does not substantially exceed what you need?
19. Operational Excellence Security Reliability Performance Efficiency Cost Optimization AWS Well Architected Framework OPS 3: How do you know that you are ready to support a workload?
20. Security a.k.a. job zero
21. Security is Automation
22. Identity Management + Encryption != Privacy
23. Live Log Analytics
24. •No patching •Improved DoS resiliency •Permission Granularity •Application Security •Dependency Security
25. Predictive Security Operations
26. Planned, Automated and Tested Incident Response
27. Reliability
28. “If you can’t feed a team with two pizzas, it’s too large” - Jeff Bezos, 2001
29. Strangling the Monolith Legacy Legacy Router Legacy Router New DNS CDN LB Zuul
30. ``
31. ”…a necessary capability if you're moving to a serious adoption of microservices.”
32. Infrastructure as Code https://aws.amazon.com/whitepapers/
34. Performance Efficiency PERF 2: How did you select your compute solution?
35. “Your database fits in RAM”
36. but also on [NVMe] disks…
37. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. NVIDIA Tesla Volta V100 GPUs “At the limit of photolithography” - Jen-Hsun Huang, Nvidia's CEO 5K FP32 GPU Cores | 15.7 Peak FP32 TFLOPS 640 Tensor Cores, 64 FMA/clock, up to 125 Tensor Core TFLOPS 16 GB GPU memory with 900 GB/sec peak GPU memory bandwidth
38. Field Programmable Gate Arrays Ryft Elasticsearch
39. Serverless
40. Handler Input/Output Types (Java) •Simple Java types (String, Integer, Boolean, Map, and List) •Plain Old Java Object •Stream type
44. “Business Logic” Message Streams Monitoring & Logging Relational Database NoSQL Build & Test Machine Learning Deployment Object Storage Migration Networking Scaling Analytics Caching
45. Host A Host B Host C Primary Secondary Replica Replica Replica Writes Reads
46. Amazon Aurora architecture Master Replica Replica Replica Availability Zone 1 Shared storage volume Availability Zone 2 Availability Zone 3 Storage nodes with SSDs § Log-structured distributed storage system designed for databases § Storage volume is striped across hundreds of storage nodes distributed over 3 different availability zones § Six copies of data, two copies in each availability zone to protect against AZ+1 failures § More performance (~5x), cost efficiency (~0.6x) § IAM Authentication, Lambda Procedures, KMS encryption and many other features SQL Transactions Caching SQL Transactions Caching SQL Transactions Caching
47. Sign up for the preview today Aurora Severless On-demand, auto-scaling database for applications with unpredictable or cyclical workloads Automatically scales capacity up and down Pay per second and only for the database capacity you use Starts up on demand and shuts down when not in use No need to provision instances NEW!
48. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. ELASTIC CONTAINER SERVICE AWS VPC networking mode Advanced task placement Deep integration with AWS services ECS CLI … { } Global footprint Powerful scheduling engines Auto scaling CloudWatch metrics Load balancers
49. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. • Cluster • Container Instances • Task Definition • Task • Service Amazon ECS Core Concepts
50. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. { "family":"hello-world", "containerDefinitions":[ { "name":"hello-world", "image":"aws_account_id.dkr.ecr.us-east- 1.amazonaws.com/hello-world", "cpu":10, "memory":500, "portMappings":[ { "containerPort":80, "hostPort":80 } ], "entryPoint":[ "/usr/sbin/apache2", "-D", "FOREGROUND" ], "essential":true } ], "volumes": [] } Amazon ECS Task Definition Container Definitions Volume Definitions
51. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. { "serviceName": "ecs-hello-world-elb", "taskDefinition": "hello-world", "loadBalancers": [ { "loadBalancerName": "EC2Contai- EcsElast-A0B1C2D3E4”, "containerName": "hello-world", "containerPort": 80 } ], "desiredCount": 10, "role": "ecsServiceRole" } Amazon ECS Service * Good for long-running applications * Load Balance traffic across containers * Automatically recover unhealthy containers * Discover services
52. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. { "serviceName": "ecs-hello-world-elb", "taskDefinition": "hello-world", "deploymentConfiguration": { "maximumPercent": 100, "minimumHealthyPercent": 50 }, "loadBalancers": [ { "loadBalancerName": "EC2Contai-EcsElast-A0B1C2D3E4”, "containerName": "hello-world", "containerPort": 80 } ], "desiredCount": 10, "role": "ecsServiceRole" } Rolling Deployments
53. Rolling Deployments minimumHealthyPercent = 50%, maximumPercent =100%
54. Rolling Deployments minimumHealthyPercent = 100%, maximumPercent = 200%
55. Monitoring with Amazon CloudWatch
56. Auto Scaling Your Amazon ECS Services
57. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Create Amazon CloudWatch alarm on a metric, e.g. MemoryReservation Configure scaling policies to increase and decrease the size of your cluster Auto Scaling Your Amazon ECS Cluster
58. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.
59. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. CHANGING COMPUTE CONSUMPTION MODEL No instances to manage Task native API Resource based pricing Simple, easy to use, powerful – and new consumption model =
60. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. RUNNING FARGATE CONTAINERS WITH ECS
61. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. RUNNING FARGATE CONTAINERS WITH ECS Use ECS APIs to launch Fargate Containers Easy migration – Run Fargate and EC2 launch type tasks in the same cluster Same Task Definition schema
62. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. TASK CPU & MEMORY CONFIGURATIONS Flexible configuration options – 50 CPU/memory configurations CPU Memory 256 (.25 vCPU) 512MB, 1GB, 2GB 512 (.5 vCPU) 1GB, 2GB, 3GB, 4GB 1024 (1 vCPU) 2GB, 3GB, 4GB, 5GB, 6GB, 7GB, 8GB 2048 (2 vCPU) Between 4GB and 16GB in 1GB increments 4096 (4 vCPU) Between 8GB and 30GB in 1GB increments
63. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Elastic Container Service for Kubernetes
64. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. It i s K u b e r n e t e s U p s t r e a m P r o d u c ti o n w o r k l o a d s S e r v i c e i n te g r a ti o n s EKS Tenets
65. EKS • Manage masters • Highly available setup • Upgrades
66. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Elastic Container Registry Fully Managed * Tight Integration with Amazon ECS * Integration with Docker Toolset * Management Console and AWS CLI Highly Available * Amazon S3 backed * Regional endpoints Secure * IAM Resource-based Policies * AWS CloudTrail Audit Logs * Images encrypted at transit and at rest
67. Does this matter to you and your customer?
68. AWS Elastic Beastalk supports: • Java • Tomcat • Docker • Multi-container docker
71. Lazy Portability Open Standards Open Source Standard Protocols Custom Driver Portability isn’t boolean
72. © 2018, Amazon Web Services, Inc. or Its Affiliates. All rights reserved. “You don’t have to be an engineer to be a racing driver, but you do have to have Mechanical Sympathy” Jackie Stewart - F1 Race Driver
73. Cost Optimization
74. Saving in content delivery with CDNs
75. Flexibility is money Reserved 3y ($ 15.91) Reserved 1y (9.87)
76. Lynn Langit - Serverless - reality or BS - notes from the trenches
77. Operational Excellence Operational Excellence Operational Excellence
80. Help everyone be well-architected! Satisficing Objective Quantitative Durable Shared
81. Julio Faerman @julioaws Obrigado!