Pitney Bowes is reinventing its business based on a SaaS and a cloud-based model to deliver services for their clients globally centered on the Pitney Bowes Commerce Cloud. The Pitney Bowes Commerce Cloud is a commerce enabler, providing access to solutions, analytics and APIs across the full commerce continuum with speed and agility to help clients identify customers, locate opportunities, enable communications, power shipping from anywhere to everywhere, and manage payments. During this session, the Pitney Bowes team will discuss the strategy behind the Commerce Cloud, how we accelerated the speed of innovation and creation of new product solutions by taking full advantage of the AWS platform, from Cloud Infrastructure Services (Compute, Big Data, Storage and Content Delivery, Databases and Networking) , to Analytics and Internet of Things Services. Pitney Bowes’ applications are deployed in Docker containers using AWS Elastic Beanstalk, and are utilizing S3, Amazon RDS, SQS, SNS, and ElastiCache. CloudWatch and CloudFormation are being used to manage the solutions that Pitney Bowes’ brings to the market. Additionally, the Pitney Bowes Data and Analytics platform is powered by Elastic Map Reduce (EMR), Spark, Aurora, DynamoDB and Postgres.
The session will also discuss Pitney Bowes’ partnership with AWS to deploy Pitney Bowes’ APIs for Location Intelligence via the AWS Marketplace, allowing developers, customers and partners to access innovative location intelligence data services in an easy to consume and highly reliable manner.
2. What to Expect from the Session
Who Pitney Bowes is and what we do to enable global
eCommerce
Why and how we moved critical Windows-based workloads
to AWS
What we are doing next to accelerate the development,
improve the quality and reliability of new eCommerce
services
3. What We Do
Anywhere to EverywhereTM
Global
Ecommerce
Digital intersects physical
to elevate and transform
the retail experience for
you and your consumers
Pitney Bowes Global Solutions
Customer Information Management
Using trusted data to drive insights and intelligence
Customer Engagement
Delivering relevant and engaging interactions
across the customer lifecycle
Shipping & Mailing
Driving parcel handling and mailing efficiency
with end-to-end innovation
Location Intelligence
Transforming location and business data into
enhanced insight
4. Pitney Bowes Transportation Network
Pitney Bowes ships parcels outbound from the US to 104
countries, and outbound from the UK to 88 countries
around the world!
5. The Life of a (simple) Parcel
5
Seller ships to PB Hub
The Hub team receives and
processes the parcel.
Domestic Linehaul Vendor
picks up the parcel
Linehaulvendor
shipstotheirhub
andprocesses.
Linehaul gets parcel
ready for
international
shipping.
Parcel transported
to departure airport.
Parcel uplifted to
destination country.
Parcel
handed
over to
customs
broker.
Broker clears parcel,
pays duty & taxes.
Parcel shipped to
consignee by local
delivery agent.
It’s here … it’s
what I always
wanted!!!
6. The Life of a (simple) Parcel
6
Seller ships to PB Hub Domestic Linehaul Vendor
picks up the parcel
Linehaulvendor
shipstotheirhub
andprocesses.
Linehaul gets parcel
ready for
international
shipping.
Parcel transported
to departure airport.
Parcel uplifted to
destination country.
Parcel
handed
over to
customs
broker.
Broker clears parcel,
pays duty & taxes.
Parcel shipped to
consignee by local
delivery agent.
It’s here … it’s
what I always
wanted!!!
The Hub team receives and
process the parcel.
Transportation Hub
software supports all
steps in this lifecycle
7. Hub Processing Overview
• Checks in each parcel
• Parcel matches order
• Parcel weight and dimensions
• Restricted and Dangerous Goods
• Open Box vs. Closed Box processing
• Parcel and Shipment documentation – Shipping Label,
Commercial Invoice, Bill of Lading, eManifest…
• Parcel Containerization
• Parcel Tracking
8. Infrastructure Situation in 2015
Aging infrastructure in our
Omaha data center
• 3-6 year-old server &
network infrastructure
• Insufficient I/O performance
• Spook factor when
investigating outages and
performance issues
9. Data Center Operations Support Change
Application
Developers
Application
Operations
NetOps
10. Data Center Operations Support Change
Application
Developers
Application
Operations
NetOps
19. Lift and Shift Selected for Transportation Hub
Lift and Shift
Amazon EC2,
AMIs, VPCs, ELBs,
Amazon Route 53
Need to get this done quickly
Development consumed with delivering
feature content
Let’s crawl before we walk
20. Development moved first
IT no longer going to manage development and QA
environments in our private data center
Dev team set up new environments in AWS in a matter of
days
Dev
Dev
21. Production Migration Key Requirements
• Seamless – no negative impact to customers and
development teams
• Improve supportability of the system
• Eliminate single points of failure (add HA)
• Improve monitoring
• Enable seasonal bursting & scaling down
• Complete Production cut over prior to holiday readiness
preparations
22. Migration Execution Step 1
• Product Services Group (operations team) built out a
“production test environment”
23. Migration Execution Step 2
• Conducted load tests
• Iteratively tuned the server configurations till they were
just right
Discovered that they needed to increased provisioned IOPS
No problem – simple config change
24. Migration Execution Step 3
Cut AMIs from tuned servers in product test environments
and used them to provision production environment
25. Migration Execution Step 4
Replicated data to AWS, configured AWS as DR site for
Omaha, and failed over during maintenance window
Omaha AWS US East
27. Results
Wire to wire it took 5 months
Performance and uptime significantly improved
• Availability Zone went down and it was just a blip on the radar
Developers and operations have much higher confidence
level in infrastructure
• The spook factor is gone
31. Evolving BorderFree To Cloud Native
• Development team is breaking up monolithic application
architecture into separate services
• Moving from Windows to Linux
• Continuous Integration environment built on AWS
• Defining AWS resources as CloudFormation Stacks
• Packaging application layer in Docker containers
• Ansible for orchestrating deployments
33. Wrap Up
• Faced with aging infrastructure
• Selected AWS primarily on cost and developer
productivity
• AWS adoption patterns – lift & shift, cloud optimized and
cloud native
• Successfully moved Transportation Hub via lift & shift,
adding high availability & true DR
• Evolving BorderFree to cloud native, building AWS
based CI with Ansible, CloudFormation & Docker