Well Archictecture Framework dotNET.pdf

© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
S U M M I T
Building well-architected .NET
apps
Imaya Kumar Jagannathan
Senior solution architect–Microsoft
platform
Amazon Web Services
S V C 2 0 9
Brian Lewis
Partner solution architect–Microsoft
platform
Amazon Web Services

“I really urge you to visit the Well-Architected
Framework because there is a lot of advice about
how you can build very extensive architectures on
top of AWS.”
Werner Vogels
Chief technology officer

S U M M I T
What is the AWS Well-Architected
Framework?
Pillars Design principles Questions

S U M M I T
Pillars of AWS Well-Architected Framework
Security Reliability
Performance
efficiency
Cost
optimization
Operational
excellence

S U M M I T
Why should I apply the AWS
Well-Architected Framework?
Build and
deploy faster
Lower or
mitigate risks
Make
informed
decisions
Learn AWS
best practices

S U M M I T
https://aws.amazon.com/well-architected/
AWS Well-Architected

S U M M I T
.NET Framework vs. .NET Core

S U M M I T
Hosting options for .NET applications
Framework
Virtual machines Containers Serverless
Serverless Lens

S U M M I T
What direction should developers go?
Answer: “It depends”

S U M M I T
Rehosting options for Windows workloads
Right size on AWS VMs
Virtual machines (VMs)
Pros
• Usually the fastest
migration strategy
• No architectural
modifications
Cons
• Miss most “cloudy” benefits
of target platform—both
technical and financial

S U M M I T
Replatforming options for Windows
workloads
AWS Well-Architected Framework
Virtual machines
(with fabric improvements)
Containers
Managed
virtual machines
Pros
• Realize some benefits of
cloud platform
• Minimal architectural
modifications
Cons
• Can be some rework
required

S U M M I T
Refactoring options for Windows workloads
AWS Well-Architected Framework
Serverless
Containers
Pros
• Realize full potential of
target cloud platform
Cons
• Requires the largest
investment and level of
effort to realize

S U M M I T
AWS Availability Zones
AZ
AZ
AZ AZ
Transit
Transit

S U M M I T
Modernize with AWS Auto Scaling
Scale your infrastructure automatically
▪ Adds or removes Amazon
Elastic Compute Cloud
(Amazon EC2) instances
based on metrics and health
checks
▪ No additional charge
▪ Minimize cost and maximize
performance
Availability Zone 2 (AZ2)
Availability Zone 1 (AZ1)
Auto Scaling Group
Minimum Size: 2
Desired Size: 4
Maximum Size: 6
AWS Auto Scaling
group
Launch
configuration
Scaling plan

S U M M I T
Multi–Availability Zone Microsoft SQL Server
Availability Zone 1
Private subnet
Availability Zone 2
AWS region
Private subnet
Synchronous commit
Automatic failover

S U M M I T
Architecture: Reference three-tier web
application

S U M M I T
Optimizing costs and performance with Windows, Intel
and AWS
• Over 57% of all Windows deployments in the cloud run on AWS, almost 2X over the closest provider
• AWS has over 10 years of experience running Windows applications
• AWS offers the broadest family of instances
• General purpose, compute and memory optimized
• Packaged, custom AMI’s
• Multiple pricing models
• Maximum performance with optimized instances
• High memory – Largest SAP certified in-memory instances
• Z1d – fastest compute with Intel Xeon Scalable processor for design automation

S U M M I T
AWS Instance Types and Intel Technologies
WS Instance Type
High Memory
instances (6 TB, 9
TB, 12 TB)
Memory Optimized
z1d
Compute
Optimized C5n
General
Purpose M5
Memory
Optimized
Burstable
I/O
Optimized
GPU GPU
Compute
P3
Memory
Optimized X1e
R5 T3 I3 G3
Intel® processor
Intel® Xeon® Platinum
8176M
Intel® Xeon® Scalable
Processors
3.0 GHz Intel Xeon
Platinum 8000
Intel® Xeon®
Platinum 8175M
series processors
Intel® Xeon®
Platinum 8000
Series
Intel® Xeon®
Scalable Proc
essors
Intel E5-2686 v4
processor
Intel Xeon E5-
2686 v4
processors
Intel® Xeo
n® E5-2686
v4
Intel® Xeon® E7 8880
processors
Intel® process technology Skylake 14nm Skylake-SP Skylake Skylake 14nm Skylake-SP Yes 22nm Ivy Bridge 14nm Broadwell
14nm
Broadwell
22nm Haswell
Intel® AVX AVX-512 AVX-512 AVX-512 AVX-512 AVX-512 AVX-512 Yes Yes Yes AVX 2.0
Intel® AES-NI Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Intel® TSX - - - - - - - - - Yes
Intel® Turbo Boost Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
SSDC+1 - - - - - - - - - Y

S U M M I T
AWS Elastic Beanstalk
Your code
HTTP server
Application server
Language interpreter
Operating system
Host
Elastic Beanstalk configures each
Amazon EC2 instance in your
environment with the components
necessary to run applications for
the selected platform. No more
worrying about logging into
instances to install and configure
you application stack.
Focus on building your application
Provided by you
Provided and managed by Elastic Beanstalk

S U M M I T
Elastic Beanstalk
AWS CodeDeploy
Amazon EC2
Replatform: Blue/green deployments on AWS
Amazon
Route 53
ELB / ALB
Auto Scaling Group
Managed AD
Multi-AZ
SQL Server RDS
Multi-AZ
ElastiCache
Auto Scaling Group
50% 50%

S U M M I T
The role of containers in modernization

S U M M I T
Windows Server containers

S U M M I T
Scheduling and orchestration
Cluster manager Placement engine
Running containers at scale with Amazon Elastic
Container Service (Amazon ECS) or Amazon ECS
for Kubernetes (Amazon EKS)
Availability Zone #1 Availability Zone #2 Availability Zone #3

S U M M I T
Hosting options for .NET applications

S U M M I T
Business benefits of modernization
Cost savings
(TCO/CO)
Resource
efficiency
Business
agility
Operational
resilience
What is it?
Launch of new products
75% faster (Unilever)
Critical workloads run in
multiple Availability Zones
and regions for robust DR
(Expedia)
50%+ reduction in TCO
(GE)
More than 500 hours per
year of server configuration
time saved (Sage)
AWS customer
success
Infrastructure cost
savings/avoidance from
moving to the cloud
Efficiency improvement by
function on a task-by-task
basis
Benefit of improving SLAs
and reducing unplanned
outage
Deploying new features/
applications faster and
reducing errors
Example Data-center consolidation Large-scale, compute-
intensive workloads
Improved agility and
developer productivity to
drive innovation and digital
transformation
IT process efficiencies
Initial
focus
Most compelling
cloud benefits

S U M M I T
Serverless
Serverless is the native architecture of the cloud that enables you to
shift more of your operational responsibilities to AWS, increasing
your agility and innovation
AWS cloud
func(object x)…

S U M M I T
AWS shared responsibility model

S U M M I T
Serverless services
• Compute – AWS Lambda , Lambda@Edge , AWS Fargate
• Storage – Amazon Simple Storage Service (Amazon S3), Amazon Elastic File System
• Data stores – Amazon DynamoDB , Amazon Aurora Serverless
• API Proxy – Amazon API Gateway
• Interprocess messaging – Amazon Simple Notification Service (Amazon SNS), Amazon
Simple Queue Service (Amazon SQS)
• Orchestration – AWS Step Functions
• Analytics – Amazon Kinesis, Amazon Athena
• Developer tooling – AWS CodeBuild, AWS CodeDeploy, AWS CodeCommit, AWS
CodeStar

S U M M I T
Well-Architecting for operational excellence
• Leverage Amazon CloudWatch and AWS
X-Ray together
• Key Lambda KPIs
• Throttling and errors
• Duration metric
• Key API Gateway KPIs
• 5XXError, 4XXError
• Integration latency
• CloudWatch custom metrics for application
insights
public string FunctionHandler(string inputString, ILambdaContext context)
{
context.Logger.Log("Application Insight KPI");
return "Success";
}

S U M M I T
Well-architecting for operational excellence
API stage variable and Lambda alias to decouple design
Map stage variable to the Lambda live alias
Alias
version1
version2
version3
version4
version5
version6
version7
version8
Stage variable = function
Prod stage
Function=version5
Beta stage
Function=version6
Test stage
Function=version7

S U M M I T
• Use AWS Serverless Application Model (AWS SAM) to package,
deploy, and model applications
• AWS SAM local for debugging Lambda function locally
• Don’t use it as a replacement for performance or regression testing
• Use Lambda Layers wherever appropriate

S U M M I T
Managing configuration variables to decouple code
from configuration
Use environment variable
When configuration is local to Lambda function
Configuration can be stored in plain text
Use AWS Systems Manager Parameter Store or AWS Secrets Manager
Configuration shared across different Lambda functions
Configuration needs to be encrypted for security

S U M M I T
Well-architecting for security
Authorize API call within API Gateway
Choose a suitable authorization method
AWS Identity and Access Management (IAM) authorization
Amazon Cognito user pools
API Gateway custom authorizer
The API keys feature is not a security mechanism IAM authorization
API Gateway custom authorizer
Amazon Cognito user pools

S U M M I T
Well-architecting for security
Track vulnerability using logs
Check compliance requirement before logging
Enable API Gateway method-level logging
Input validation
Use API Gateway basic request validation
Implement application-specific deep validation

S U M M I T
Well-architecting for reliability
Monitor and manage service
limits
Ensure significant gap between service
limit and hard limit
Isolate Lambda concurrent execution limit
for critical and non-critical paths
Workloads across different accounts
based on profile, threat, and org structure
Regulate access rates
Enable API-level throttling
Control usage using API keys

S U M M I T
Well-architecting for reliability
Asynchronous calls and events
Follow asynchronous pattern where possible
Define SLA and fail-in-flight requests overriding it
Enable dead-letter queues (DLQs) for fails
Build for resiliency
Prefer Step Functions where possible
Implement stage pattern to rollback transactions

S U M M I T
Well-architecting for performance efficiency
Leverage optimization settings
Caching in API Gateway
DAX in DynamoDB
Global and local secondary indexes in
DynamoDB
Global scope for Lambda function

S U M M I T
Well-architecting for performance efficiency
• Deploy with consideration
• Deploy Lambda in an Amazon Virtual Private Cloud
(Amazon VPC) only when necessary
• Plan multi-Availability Zone NAT gateways
• Pre-JIT your Lambda code
• Use Amazon Linux image to compile the solution

S U M M I T
Well-architecting for cost optimization
• Resource allocation
• Load test for optimal memory allocation
• Optimize CloudWatch logs
• Configure log level such as INFO, DEBUG
• Control log size using retention period
• Optimize execution
• Avoid unnecessary Lambda functions
• Optimize code execution time
• Let Step Functions handle orchestration
• API Gateway validations also help reduce
costs
Call service using Lambda
Call service using service proxy
Amazon Kinesis
Data Firehose
Kinesis Data
Firehose
Amazon
S3
Amazon
S3
Amazon S3
Bucket for
Log Files

S U M M I T
Case study

Thank you!
Imaya Kumar Jagannathan
@imaya
Brian Lewis
@BrianLewis_

Well Archictecture Framework dotNET.pdf

Recommended

More Related Content

Recently uploaded (19)

Featured (20)

Well Archictecture Framework dotNET.pdf