Salesforce and MuleSoft: six integration engagements in one study
For several years Stratoflow was a Salesforce and MuleSoft development partner. We no longer sell that as a standalone service, but the work is why our integration layers are designed by people who have run a commercial integration platform in anger. Six engagements, one page.
- 6
- engagements: two for platform vendors, four for their customers2016 to 2023
- Thousands
- of customised orders a day routed by the ecommerce integration platformProduction
- Incremental
- Salesforce-to-AWS data loads feeding machine learning, only new records movedProduction
- Clients
- MuleSoft; Otwarte Klatki (Anima International); a San Francisco machine-learning startup; a healthcare IoT startup; a European clothing customisation retailer; a Salesforce customer building predictive models
- Platforms
- Salesforce Platform, MuleSoft Anypoint Platform and Mule Runtime, Amazon AWS (S3, Elastic MapReduce), Eclipse-based Anypoint Studio
- Stack
- Java, MuleSoft connectors and custom connectors, RAML, LSP4J language servers, Salesforce web services and workflows, SFTP, FTPS, EDI, CSV and XML exchange
- Offer today
- Integration work inside Application Modernization Sprints and Performance Rescue; Technical Due Diligence on existing integrations
A RAML plugin for MuleSoft's Anypoint Studio
MuleSoft engaged Stratoflow for specialised Java work: a RAML (RESTful API Modeling Language) plugin for the Eclipse-based Anypoint Studio, so developers could design and test APIs without leaving the tool. We designed the plugin to fit MuleSoft's existing ecosystem and standards, coordinated the deliverables, and implemented and tested it.
The plugin implemented the Language Server Protocol with LSP4J, giving RAML authors syntax highlighting, auto-completion and error detection inside Anypoint Studio. Developers could design, test and iterate RAML APIs in place, shortening their development cycle.
Salesforce as a general-purpose data store, generated from RAML
A Salesforce customer wanted to enrich its records with context from many external systems, mostly documented in RAML, for machine-learning analysis without heavy prefiltering. Using MuleSoft we built a universal mechanism that generates a custom Salesforce data model from an API's RAML specification and then populates it, as a full snapshot or incrementally.
The tool turned Salesforce into a general-purpose data store and let the business try new data sources for its predictive models in days rather than projects.
Machine learning between Salesforce and Amazon AWS
A San Francisco startup building a custom machine-learning system on Salesforce data needed high-volume data moved regularly to AWS. We built a MuleSoft integration that identifies new Salesforce data, applies the transformations and moves only the new chunks to Amazon S3 as the load area for Elastic MapReduce jobs, deployed in the cloud with scaling up and down by input volume.
Incremental loads kept the model and its reports current while limiting the resources and cloud cost that full snapshots would have consumed.
A healthcare IoT SaaS on the Salesforce platform
A healthcare startup needed a secure, scalable cloud platform for a SaaS connected to Internet of Things devices in the homes of elderly people, monitoring critical activities. The drivers were scale-out, pay-per-use, high availability, security, compliance that large partners and resellers would accept without concern, and business processes that non-technical users could change quickly.
Salesforce Platform became the central development and database platform. Workflows orchestrated sales and order processing including physical device delivery; custom web services integrated delivery, accounting and logistics partners; custom pages provided status and reporting. Security and healthcare compliance came with the platform on day one, and the monitoring logic stayed editable by the product owner and analysts.
Salesforce for Otwarte Klatki, part of Anima International
The Polish NGO Otwarte Klatki (Open Cages) ran fundraising on fragmented systems that created inefficiency and administrative load. We implemented Salesforce, developed the data model and built the CRM on it, migrated data from Excel, integrated the donation system, payment, donation and store operations, automated processes across the organisation and trained the team.
Fundraising was strengthened, including regular monthly giving; data management, auditing and infrastructure improved; and volunteers' administrative workload fell, leaving more time for the organisation's mission.
An ecommerce platform integrated with fabric label printing and production partners
A clothing customisation service operating across several European countries revamped its ecommerce platform with us. Each highly customised order goes to a different production partner, and some partners imposed communication protocols that forced obsolete integration approaches.
A flexible integration toolkit let us add and amend integrations by configuration. Connectors covered SFTP, FTPS, DBLink over the internet, EDI, CSV and XML as well as REST, with custom integrations implemented as connectors to the same framework. The platform now processes and routes thousands of orders a day and became the foundation for the client's omnichannel architecture.
They provided superb service with seamless communication and a highly professional, technical approach. The team displays impressive technical expertise and are willing to share information and engage in constructive feedback.
Filip StachnikOperations Manager, Otwarte Klatki (part of Anima International)Integration platforms taught us what integration layers cost
We no longer sell Salesforce or MuleSoft development as an offer. What stayed is the discipline: decide which system owns each fact, design for the partner that times out, make protocol a configuration detail, and measure the integration layer as carefully as the systems it connects. When an integration problem lands inside a Java system today, the same engineers handle it with MuleSoft, Apache Camel or plain Spring Integration, whichever your systems already use.
Where this work lives now
Book a scoping call
Thirty minutes with an architect, not a salesperson. You leave with a written view of scope, price model and whether we are the right team for it.