2. Welcome
Sheri Sullivan
Senior Marketing Manager
Global SI Ecosystem
Amazon Web Services
3. Webinar Overview
• Submit Your Questions using the Q/A tool.
• A copy of today’s presentation will be made available on:
• AWS SlideShare Channel@ http://www.slideshare.net/AmazonWebServices/
• AWS YouTube Channel@ http://www.youtube.com/user/AmazonWebServices
Special Note: Today’s Webinar is being recorded.
4. What We’ll Cover
• AWS Cloud Computing
Overview
• eCommerce Platform
Infrastructure
• Preparing for your
eCommerce business for
peak season
5. E-COMMERCE
WITH AMAZON WEB SERVICES
Simone Brunozzi
Technology Evangelist
Amazon Web Services
simone@amazon.com
18. CERTIFICATIONS
& COMPLIANCE
SAS 70 Type II audit
ISO 27001 Certification
Sarbanes-Oxley SOX
HIPAA healthcare
FISMA US Federal Government
DIACAP MAC III Sensitive IATO
21. On a global footprint
Region
US-WEST (N. California) EU-WEST (Ireland)
GOV CLOUD ASIA PAC
(Tokyo)
US-EAST (Virginia)
US-WEST (Oregon)
ASIA PAC
(Singapore)
SOUTH AMERICA (Sao
Paulo)
23. On a global footprint
Edge Locations
London(2)
Seattle South Bend New York (2) Amsterdam
Newark Stockholm
Dublin
Palo Alto
Tokyo
San Jose
Frankfurt(2)
Paris(2)
Ashburn(2) Milan
Osaka
Los Angeles (2) Jacksonville
Dallas(2) Hong Kong
St.Louis
Miami Singapore(2)
Sydney
Sao Paulo
30. DEMOCRATIZE SEARCH
Amazon CloudSearch delivers a fully-managed search
service in the cloud that can be set up and running in
less than 1 hour, with automatic scaling for data &
traffic, at a price starting at less than $100 per month.
46. The Way Consumers Buy is Changing
Consumer purchase behaviors—online and in-store still dominate, but…
Safest Most reliable Overall favorite Easiest Most convenient
77
%
69 68 68
% % %
59
%
38
31 %
28 % 27
% %
22 20
% %
13 13
11 % %
7% %
In-store Purchase Online PC Purchase Mobile Purchase
*Source: Nielsen
46
47. Mobile is an increasingly important channel —
tablets in particular
iPads take the lead:
• Account for 2/3 of shopping, browsing and purchasing
• 90% of mobile revenues (creeping ahead of desktop orders)
47
48. Running eCommerce in the Cloud
• Applications must be architected for the cloud
• Decouple components
• “Share nothing” architectures
• Consider mobile, social APIs
49. Best Practice #1
Horizontal Scalability vs. Vertical Scalability
• Proactive scaling
• Anticipated traffic (peaks)
• Marketing-driven traffic
• Reactive scaling
50. Best Practice #2
Performance Is a Business Requirement
• Correlation between response time and conversion
• Cache at different levels
• Build representative test cases
• Measure performance metrics
• Server & client performance
51. Best Practice #3
Automate EVERYTHING
• Infrastructure as code
• Configuration & environment management
• Server provisioning time
• No human error
52. Best Practice #4
Do Your Capacity Planning
• Pay for what you use
• Err on the side of extra capacity
• Mix planned & elastic capacity
• Plan ahead
53. Best Practice #5
Managing Environments
• AWS enables
• Limited capital investment
• Short time period
• Development, systems integration, testing
• Efficient creation and removal of intermediate
environments (QA, Test Harnesses)
55. Compliance & Security
• PCI
• Security
• Burden is on the
merchant
• Level 1, Level 2
compliance
56. Beware of Licensing
• Understand software
licensing restrictions in
cloud environments
• Work with your
software provider
57. What can you do now?
Testing
• Spin up new QA environments quickly for parallel release testing or
overlapping releases
• Create Load Testing environments as large or larger than Production
Experimentation
• Test new features during code freeze
• Quickly setup A/B test beds
• Social network features outside of true production environment
58. What can you do now?
Production Capacity Overflow
• Offload portions of production traffic to servers in the
cloud such as heavy read-only content, CMS or
catalogue
• Create gateways for unpredictable traffic such as
mobile APIs or social network integrations
59. Optaros’ View of the Digital Commerce Ecosystem
CUSTOMER EXPERIENCE
STOREFRONT MOBILE SHOPPING COMMUNITY MARKETPLACE
CART
ENABLING TECHNOLOGIES
Digital
Commerce
PRODUCT
CATALOG
PRODUCT CONTENT
MANAGEMENT
ANALYTICS Enablement
INTEGRATION
ORDER ERP CRM PAYMENT CUSTOMER FULFILLMENT
MANAGEMENT GATEWAY SERVICE
60. About Optaros
Optaros delivers digital
commerce services at an USER EXPERIENCE
accelerated rate.
TECHNOLOGY
We have the capabilities to
provide award winning user
VISUAL DESIGN
experiences & creative
design, and the technology skills
needed to create reliable and HOSTING & SUPPORT
scalable technology
infrastructure.
We are often asked the question: how did Amazon get into cloud computing? Amazon is really good at providing an immense selection of products, and of shipping those products to customers efficiently. But behind that online capability lies years of experience in providing technical services to the business that ensures our online stores are secure, fast, always available and capable of meeting huge seasonal demand.
To help understand why Amazon Web Services and Cloud Computing are changing IT delivery, a nice comparison to make is that of a utility like electricity. When electricity was discovered businesses would generate their own, using steam generators to power factories. When electricity was brought together under a national system of supply, it was no longer necessary for everyone to generate their own and buy and maintain their generators, you could simply tap into the grid and use what you needed, paying only for what you did use, and be assured that the electricity you consumed was consistent and always available.
And scale is something AWS is used to dealing with. The Amazon Simple Storage Service, S3, recently passed 1 trillion objects in storage, with a peak transaction rate of 750 thousand per second. That's a lot of objects, all stored with 11 9's of durability.
Autreproprieteinteressanted'AWSestqu'on ne paye pluslorsqu'oneteint les resources …Avec infra traditionnelle, an eteint le serveur, on economise seulementsurelectricite ;-)
PCI DSS Level 1AWS satisfies the requirements under PCI DSS for shared hosting providers. AWS also has been successfully validated against standards applicable to a Level 1 service provider under PCI DSS Version 2.0. Merchants and other PCI service providers can use the AWS PCI-compliant technology infrastructure for storing, processing, and transmitting credit card information in the cloud, as long as those customers create PCI compliance for their part of the shared environment. Amazon Elastic Compute Cloud (EC2), Amazon Simple Storage Service (S3), Amazon Elastic Block Storage (EBS) and Amazon Virtual Private Cloud (VPC) were included as part of this validation. Under the same circumstances, other enterprises can also benefit by running their applications on other PCI-compliant technology infrastructure. AWS provides additional information and frequently asked questions about its PCI compliance on its web site.
And just like an electricity grid, where you would not wire every factory to the same power station, the AWS infrastructure is global, with multiple regions around the globe from which services are available. This means you have control over things like where you applications run, where you data is stored, and where best to serve your customers from.
Each AWS region is also split into Availability Zones, making highly available applications possible from within a region.
And the whole footprint is supported by many edge locations, places from which content can be served to your customers for the fast possible response times.
cloud-hosted search service from AWSfully managed search service on the cloudscales automaticallyeliminates complex managementsupport structured or unstructured text documents. Why ?huge explosion on amount of data created everyday. structured or not.
Each autoscaling group uses a different set of AMIAMI-B contains the test code.
Pour illustrer le launch early
Last, but not least, with Cloud Computing allows you to focus more time on your business.
How does Amazon talk about this slide?
42% of mobile users browsed products, 32% read online reviews, and 23% purchased products on mobile in the last 30 days (Nielsen Smartphone Analytics December 2011)http://allthingsd.com/20120604/e-commerce-accelerating-due-to-personalization-pinterest-and-ipad/2012 Q1 Mobile Study sponsored by RichRelevance, mobile’s share of e-commerce grew nearly 250 percent in the year ended March 31, 2012 — and the iPad accounts for two-thirds of the shopping, browsing and purchasing in this emerging channel. 90 percent of all mobile revenue is generated via iPads. The iPad’s average order value is even creeping ahead of desktop-based orders.
- Consider vertical scalability in last resort (difficult to go completely without a relational/transactional db in eCommerce though) - Scale with demand- Proactive scaling based on typical traffic shape (peak hours, peak days) - Proactive scaling based on marketing schedule (new product launch, private sales event, cyber monday campaign) - Reactive based on alerts (sever load, …)- Licensing model can be an issue there depending on your ecommerce platform (lack of "pay as you go" model) - Performance IS a business requirement - There is a direct correlation between response time and conversion rate- Cache at different level (full page, blocks/server side includes, ajax, localStorage, …) - Build test cases representative of expected load - and run frequent performance test and optimization campaign - Continuously measure performance metrics in production and customer satisfaction (APDEX)- Consider both server and client side performance - Automate Everything- Infrastructure as code (chef/puppet/...)- Configuration and environment management is key- Server provisioning time (maintain your AMI with key software)- Avoid human errors - automate admin action through scripts and interfaces- PCI and Security in the Cloud- Limit the scope of PCI - ideally exclude your commerce platform from the PCI scope- Most of the burden is on the Merchant and is operational. This is usually a important investment and they are some ways to reduce the scope to a very manageable part- Achieving and Maintaining PCI certification on AWS is not much different (costly/complex) than any PCI 2.0 DSS certified hosting provider- Sensitive customer application can be encrypted- Plan for failure - or how to design resilient applications in the cloud- Inherently to commodity hardware and shared environment, the individual server availability is usually lower.- Amazon does not provide server level SLA, Amazon does not provide availability zone SLA - 99.95% is for at least 2 AZ- Key elements to take into account in infrastructure design - Leverage multiple availability zones (spread front-end servers across zone, RDS Multi-AZ, Elastic Load Balancer)- Elastic IPs to remap failed server- Monitor your infrastructure as well as (CloudWatch and others tools)- S3 and EBS storage allow to store data outside of instances- Avoid SPOF as possible - Thirds party dependencies- Disaster Recovery (another AWS Region disaster recovery plan cover most cases, but some multi-cloud options can be considered) - Cloud "lock-in" - Some Amazon services are more portable than other to other cloud providers - EC2 (Linux/Virtualization), S3/EBS (Storage), RDS (Mysql), Elastic Cache (Memcached), SES (smtp), EMR (Hadoop), ELB (most LB), CloudFront(most CDN)- Some are more tight to Amazon Intellectual Property and while they provide a lot of value they create a stronger dependency - DynamoDB (versus another NoSQL), SQS (versus another message queue), CloudSearch (versus another search engine) - This is an important design decision to select what service are best to be leveraged - Capacity planning - Pay for what you use and adapt expenses to demand- For shot period of time it is better to have a bit of extra capacity to support unexpected demand, and optimize after, that to fail due to capacity limit - Mix planned capacity (reserved instances) and elastic capacity (on-demand) to lower costs (there might be a better diagram where we overlay the reserved capacity on that diagram) - Mobile Commerce - tbc - Environments Management - Test Harness : Historically having a QA/pre-prod environment similar in complexity and size has been impossible. - AWS allows your do achieve that by creating with limited capital investment environment representative to your production environment - Possible to run for a short period of time several environment version for new releases and lower maintenance windows size for deployments without increasing cost - Possible to run development and system integration and testing environment on similar infrastructure to reduce risks of bugs not identified earlier due to environment differences - Architecture for the Cloud - AWS reference architecture for web-based solutions (applies well to e-commerce solution)( We could make a variation of our typical ecommerce reference architecture including for example search engine, cache servers …)
Kevin has examples of marketing-driven traffic- Performance IS a business requirement - There is a direct correlation between response time and conversion rate- Cache at different level (full page, blocks/server side includes, ajax, localStorage, …) - Build test cases representative of expected load - and run frequent performance test and optimization campaign - Continuously measure performance metrics in production and customer satisfaction (APDEX)- Consider both server and client side performance - Automate Everything- Infrastructure as code (chef/puppet/...)- Configuration and environment management is key- Server provisioning time (maintain your AMI with key software)- Avoid human errors - automate admin action through scripts and interfaces- PCI and Security in the Cloud- Limit the scope of PCI - ideally exclude your commerce platform from the PCI scope- Most of the burden is on the Merchant and is operational. This is usually a important investment and they are some ways to reduce the scope to a very manageable part- Achieving and Maintaining PCI certification on AWS is not much different (costly/complex) than any PCI 2.0 DSS certified hosting provider- Sensitive customer application can be encrypted- Plan for failure - or how to design resilient applications in the cloud- Inherently to commodity hardware and shared environment, the individual server availability is usually lower.- Amazon does not provide server level SLA, Amazon does not provide availability zone SLA - 99.95% is for at least 2 AZ- Key elements to take into account in infrastructure design - Leverage multiple availability zones (spread front-end servers across zone, RDS Multi-AZ, Elastic Load Balancer)- Elastic IPs to remap failed server- Monitor your infrastructure as well as (CloudWatch and others tools)- S3 and EBS storage allow to store data outside of instances- Avoid SPOF as possible - Thirds party dependencies- Disaster Recovery (another AWS Region disaster recovery plan cover most cases, but some multi-cloud options can be considered) - Cloud "lock-in" - Some Amazon services are more portable than other to other cloud providers - EC2 (Linux/Virtualization), S3/EBS (Storage), RDS (Mysql), Elastic Cache (Memcached), SES (smtp), EMR (Hadoop), ELB (most LB), CloudFront(most CDN)- Some are more tight to Amazon Intellectual Property and while they provide a lot of value they create a stronger dependency - DynamoDB (versus another NoSQL), SQS (versus another message queue), CloudSearch (versus another search engine) - This is an important design decision to select what service are best to be leveraged - Capacity planning - Pay for what you use and adapt expenses to demand- For shot period of time it is better to have a bit of extra capacity to support unexpected demand, and optimize after, that to fail due to capacity limit - Mix planned capacity (reserved instances) and elastic capacity (on-demand) to lower costs (there might be a better diagram where we overlay the reserved capacity on that diagram) - Mobile Commerce - tbc - Environments Management - Test Harness : Historically having a QA/pre-prod environment similar in complexity and size has been impossible. - AWS allows your do achieve that by creating with limited capital investment environment representative to your production environment - Possible to run for a short period of time several environment version for new releases and lower maintenance windows size for deployments without increasing cost - Possible to run development and system integration and testing environment on similar infrastructure to reduce risks of bugs not identified earlier due to environment differences - Architecture for the Cloud - AWS reference architecture for web-based solutions (applies well to e-commerce solution)( We could make a variation of our typical ecommerce reference architecture including for example search engine, cache servers …)
- Capacity planning - Pay for what you use and adapt expenses to demand- For shot period of time it is better to have a bit of extra capacity to support unexpected demand, and optimize after, that to fail due to capacity limit - Mix planned capacity (reserved instances) and elastic capacity (on-demand) to lower costs
- Environments Management - Test Harness : Historically having a QA/pre-prod environment similar in complexity and size has been impossible. - AWS allows your do achieve that by creating with limited capital investment environment representative to your production environment - Possible to run for a short period of time several environment version for new releases and lower maintenance windows size for deployments without increasing cost - Possible to run development and system integration and testing environment on similar infrastructure to reduce risks of bugs not identified earlier due to environment differences
(there might be a better diagram where we overlay the reserved capacity on that diagram)
- Mobile Commerce - tbc
- Mobile Commerce - tbc
The technology components that enable commmerce include the front end, the back end, and the middle office—and the tools and technologies that drive each of these—and that require complex integrations with one another.This ecosystem can be challenging to understand and support because it is at the intersection of Strategy, Creativity and Technology and traditional agency and SI partners have not been able to cross between these issues well The consumerization of back end IT is another disruptive change influencing the strategic direction of many large corporations and it crosses many traditional boundaries