Building a Resilient Architecture for Cross Border Ecommerce and Last Mile
The transportation and logistics (T&L) industry covers a wide range of services, such as multi-modal transportation, warehousing, fulfillment, freight forwarding, and delivery. Each logistics service provider offers a selection of these services, for instance specializing on a specific leg of the global supply chain or on specific value-adding functions. This specialization makes every company unique in its own way and defines its core technologies and digital vision.
Customers looking to start or expand their cross border ecommerce service (one of the T&L growing products) request solutions that enable a faster shipment delivery, supply chain visibility, cognitive customer experience (customer interaction augmented by AI), and operations efficiency.
Last mile ecommerce, which is the last leg in ecommerce shipment lifecycle, has a huge potential to benefit from the AWS Cloud and the emerging technologies offered today.
The template architecture (described in the following sections) answers the common core capabilities built into the vision of each ecommerce and last mile product.
Cross border ecommerce is a mix of B2C, B2B, and B2B2C services. Customers upend their operating models to unleash revenue growth across all these models integrating different styles of ecommerce. As customers expand their business by opening new origins and destinations of service, the architecture proposed is scalable to include all of these models and the global nature of the solution.
A typical T&L core will cover the shipment alerting, consolidation, fulfillment, and tracking. Those core operational services are integrated with a core billing and costing engine to create a shipment billing document that gets integrated with the blackened ERP for invoicing.
The alerting services create the shipment/order and captures the first structure of an airway bill (AWB) that gets updated as the shipment moves in its journey. A shipment record has all the shipment details like weight and commodity but also captures the shipper and consignee and billing party of that shipment. The alerting services are invoked by different sources from EDI, APIs, ecommerce plugins, operations apps, and portals.
Consolidation/manifesting or bagging, as some customers might prefer to call it, is a core service where shipments traveling to the same destinations are grouped and attached to a console MAWB (master air way bill).
Tracking and Event Logging
Every ecommerce shipment is appended to tens of tracking updates on average. Those tracking updates are vital to fire events through the shipment operational or billing lifecycles. They are also the base for the customer and consumer tracking experience through the different touchpoints.
Online tracking services listen to all the different events happening on a shipment like shipment created, manifested, delivered, lost, held in customs, paid, returned, and many others.
The order of events capturing and retrieval in a near-live streaming is a core capability. IoT connectivity for ecommerce has many uses cases from capturing the location and tracking updates to logging the shipment shipping surroundings like temperature and humidity where needed.
T&L providers operate through a large network of partners. The ecosystem of partners helps complete and fulfill the end-to-end shipment lifecycle.
So, whether it’s a first leg partner injecting business into the customer’s network or a last mile partner streaming the whereabouts of the shipment, tracking details, or IoT readings, a partner’s quick onboarding is a core design capability.
Operating in a multi-partner structure mandates flexibility to cater for different modes and structures of integration such as different EDI templates, different API structures, and mappings.
Machine Learning to Drive Business Decisions
Last mile operations has many challenges because many companies still employ many manual processes in their operations.
Business process efficiency is constrained by low margins and by capacity limitation, which is highly impacted by the volatility and fluctuation of the business.
This calls for machine learning and automation to cut down on the manual work. Route optimization, dynamic routing, demand prediction, capacity planning, transit time prediction,
Ecommerce consumers are looking for a flawless digital experience through the different direct customer touchpoints like customer service teams and ground couriers or indirect digital touchpoints like portals, consumer apps, and bots.
B2C companies give their customers visibility to their shipments’ status, flexibility to customize their delivery experience, updates about the transit time delivery schedule, to give instructions and pay online in a user-friendly responsive design.
Reaching out to consumers through the marketing campaigns and social media, notifications and customer service tools like outbound calling, bots, or virtual assistants are also relevant to the industry and products like Amazon Pinpoint and Amazon Connect are part of the architecture.
Based on the ecommerce services the consumer opts into, they might be allowed different payment options. Online payment portals, cash on delivery, PUDO check options, duty calculators, and customs integrations are in place.
Integrating with payment gateways or smart customer wallets are required.
End-to-end Ecommerce Reference Architecture
High-level Value Proposition
The purpose of this architecture (figure 1) is to show the art of the possible in terms of functionalities, integration, and best practices. It won’t be necessarily to implement the fully scaled architecture for a customer, especially during the first phases, so risk, complexity, and costs will be reduced. At the same time, the microservice-oriented design will allow to expand the scope and capabilities without significantly impacting the existing modules.
This architecture is predominantly serverless, with AWS Lambda and Amazon Elastic Container Service (Amazon ECS) providing flexible, scalable, and cost-efficient compute power respectively at the microservices- and code-execution level. AWS takes the heavy lift of infrastructure and resources provisioning with majority of the services shown being fully managed, such as AWS IoT Core, AWS Lambda, Amazon Simple Queue Service (Amazon SQS), Amazon ECS, and Amazon S3. This reduces overheads from undifferentiated activities and instead enables innovation at a faster clip through improved DevOps tempo.
Compliance with Well-Architected Pillars
This architecture is built in line with the AWS Well-Architected Framework. Specifically:
- Enables operational excellence: each component of the architecture and related microservices is run and monitored through a shared services layer inclusive of Amazon CloudWatch, AWS Systems Manager, and AWS Key Management Service (AWS KMS). This enables for automating changes, responding to events, and defining standards to manage daily operations.
- Provides state-of-the-art protection of information and systems: AWS Identity and Access Management (IAM) and AWS KMS confidentiality and integrity of data, identifying and managing who can do what with permission management. All the AWS services and underlying software and hardware infrastructures are compliant with most stringent industry, governmental, and military data security standards ensuring highest grade of systems protection and establishing controls to detect security events.
- Maximizes performance efficiency of both IT and computing resources: fully managed services and a containerized architecture always enable to provision the right resource types and sizes based on workload requirements, monitoring performance, and making informed decisions to maintain efficiency as business needs evolves. This high-level reference architecture will be further refined as we progress in the solution effort to optimize performances both holistically and at microservice level.
- Optimizes costs and allows close cost monitoring: AWS has a set of solutions to help you with cost management and optimization. This includes services, tools, and resources to organize and track cost and usage data, enhance control through consolidated billing and access permission, enable better planning through budgeting and forecasts, and further lower cost with resources and pricing optimizations. For instance, Amazon CloudWatch enables the most granular observation and control over cloud usage.
- Provides a highly reliable framework: this architecture has a distributed system design, with robust disaster recovery setup. The serverless, fully managed design of its components allows for automated recovery from failure, automated test recovery procedures, horizontal scalability to increase aggregate workload availability and automation of change management.
Functional Description of the Architecture’s Layers (Left to Right)
Ingestion layer: the architecture allows ingestions of data from multiple sources using different technologies and protocols, such as:
- Customer’s internal systems, databases, ERPs via API
- External partners ERPs via API, EDI, SFTP
- Desktop and mobile booking portals via API
- Third-party data providers (weather, traffic, customs) via both API and EDI
- IoT sensors via MQTT and other most common protocols
Partners connectivity and management: it is possible to implement a third-party connectivity and management service, either containerized or running on EC2 instances (pending on the solution of choice) that will be provided by AWS or AWS Partners with state-of-the-art competency in EDI management.
The Shared Service layer provides services monitoring, notifications, identity and access management functions, system management, and keys management.
The Customer Experience layer allows integration of email services, a fully scaled digital contact center experience with Amazon Connect, and the bulk management of marketing communication with Amazon Pinpoint. Optionally, a graph database like Amazon Neptune can support advanced customers relationship analysis workloads.
Containerized apps for supply-chain management: this architecture includes microservices dedicated to specific supply chain management functions. Booking and receiving, consolidation, global tracking, fulfillment, and delivery planning functions will be launched by applications running on an Amazon ECS framework. Specific applications can be selected by the customer among the ones offered by AWS Partners. AWS services and infrastructure will provide data synchronization, events propagation, and workflows automation between the different applications.
Database layer: AWS can provide a diversified range of databases, each one addressing a specific business need. Instead of legacy SQL database with punitive licensing or open source framework with higher implementation complexity, Amazon Aurora is a fully managed MySQL and Postgres compatible service that has several-times-faster performance than the typical high-end implementations. Amazon DynamoDB is an extremely fast key-value database, ideal for recording high throughput/high volume data flows associated with postal and parcel workflows. Amazon Timestream is ideal for data series that changes over time and can support for instance cost-optimization analysis across shipments’ lifecycle. Finally, Amazon Quantum Ledger Database (Amazon QLDB) offers a ledger database that can be used to implement blockchain-based track-and trace features across supply chains. Amazon S3 is the storage service that will allow the creation of data lake where raw data can be stored, deep-stored and staged, establishing a cost-effective and secure source of truth for the entireorganization.
Shared capabilities. This is a client layer where powerful analytics will be enabled leveraging on the underlying database and data lake layer. Functions performed will include:
- Data warehousing with Amazon Redshift
- Real-time dissemination of alerts and milestones to internal and external stakeholders with Amazon SNS
- Business intelligence dashboards embeddable into existing websites/applications or consumable as standalone resources provided through Amazon QuickSight
- Predictive and prescriptive analytics with the full machine learning stack managed by Amazon SageMaker
AP/AR billing. Amazon EventBridge is an integration bus enabling connectivity between SaaS application and AWS services. In addition, the more specific AWS Connector for SAP can be implemented. This will allow integration with most common financial ERPs enabling this architecture to be compatible with existing financial/accounting workloads without the need of large scale adjustment to technology, people training, and processes.
AWS offers the building blocks, industry expertise, and technical competencies to enable customers running complex ecommerce operations seamlessly and more cost effectively, even when supply chain spans from the global to the last miles. The breadth and depth of services, in conjunction with the most vibrant community of technology and integration Partners allow to both embed existing on-premises workloads into new cloud architectures and develop cloud-native solutions. To learn more please visit AWS for Industrial.