Home · Case studies

Case studies

3 engagements, in detail

Short reference lines prove nothing. These three are written out in full: what the client had, what broke, what we built, and the figures the platforms actually run at today. Clients are not named, at their request and ours.

What you will find here

21 years, several hundred projects, 3 platforms still evolving

Each of these platforms has been through several technology generations since we built it: message buses replaced, event streaming introduced, API management and identity rebuilt, cloud migration under way, a next-generation platform in build alongside production. The figures below are what they run at today, not what they were specified for.

Global container shipping

A custom ESB that grew from 9 messages a second to 2 million exchanges a day

Customs declarations, container movements, invoicing, vessel tracking: every one of these operations is an integration flow. When they stop, the freight stops. We designed the platform that carries them and we have run it ever since.

One of the world's largest container shipping and logistics groups · Partnership since 2013 · platform designed in 2017 · next generation in build for 2026

Global container shipping
~2MExchanges per dayUp from 9 messages per second at launch, with no rewrite of the core architecture.
99.97%Measured availabilityContinuous uptime measurement over the life of the platform.
Top 5Most stable applicationsRanked in the group's five best applications for performance and stability, across the whole portfolio.
10+ yrsContinuous partnershipRenewed without interruption since 2013.

Context

This group moves tens of thousands of containers around the world every day. Customs authorities, port authorities, carriers, suppliers, customers and the vessels themselves all exchange structured data with the group's systems, continuously and in several formats. Those flows are the nervous system of the operation rather than a back-office concern.

The relationship began in 2013. In 2017, faced with high-cadence EDIFACT exchange requirements, we designed and deployed a custom enterprise service bus as the backbone of the group's data integration. 10 years later it carries close to 2 million exchanges a day and remains the reference architecture, with no major rewrite of its technical core.

What made it hard

  • High-cadence EDIFACT from day one9 messages per second required at launch, through a database to XML to EDIFACT chain, with a wide variety of formats and message-specific business rules.
  • Designing for volumes nobody had specified yetThe architecture had to absorb a rise from 9 messages per second to 2 million daily exchanges over 10 years, along with new formats and business uses that did not exist at design time, without touching the application core.
  • Continuity in a 24/7 critical environmentMaritime freight runs continuously. The platform tolerates no interruption, which sets the bar for both stability and resilience.
  • A fully remote team, for 10 yearsContinuity, knowledge transfer and the absence of any single point of failure had to be engineered into the team as deliberately as into the platform.

What we did

Custom ESB architecture

A proprietary ESB designed from the outset to absorb long-term growth in volumes and business needs. Built around EDIFACT and XML flows, and kept flexible enough to take on new uses without a core rewrite for more than 10 years. The difference was not the tooling but the architectural choices made at design time.

  • Custom ESB
  • EDIFACT
  • XML
  • Content-based routing
  • Transformation
A remit deliberately wider than the contract

The contract covered transformation and routing. What we actually delivered also includes the user interface layer, deployment automation, industrialised development processes and end-to-end functional development. The team identifies needs before they arrive as tickets. That posture is the main reason the partnership has lasted.

  • CI/CD
  • Industrialisation
  • UI
  • Business features
A resilient, industrialised organisation

A fully remote team of ten, four of them ours forming the technical core. Every task documented in JIRA and Confluence and transferable at any moment. An automation-first house method: developers develop, they do not administer. No single point of failure anywhere on the project.

  • JIRA
  • Confluence
  • Remote
  • Automation-first

What comes next

A next-generation platform was approved at the end of 2024 and is being built alongside the running system rather than replacing it in one step, to improve performance, stability and operational efficiency without breaking the existing foundation. Target delivery: 2026.

Technical stack

  • Custom ESB
  • EDIFACT
  • XML
  • EDI message routing
  • Automated deployment pipelines
  • JIRA
  • Confluence

The offers behind this engagement

European public financial institution

From point-to-point spaghetti to an ESB, without a single maintenance window

Integration architecture, middleware, data management and platform administration for an institution that processes millions of transactions, files regulatory reporting and exchanges cross-border financial messages in real time.

A European public financial institution · Partnership since 2007 · middleware, event streaming, API management and data virtualisation, each renewed in turn

European public financial institution
8MAPI calls a dayHandled daily with no service interruption, continuous real-time monitoring and automatic failover.
×12Faster processingCritical business processes cut from 60 minutes to 5. Business teams moved to near-real-time data.
0DowntimeWeekly maintenance windows eliminated for good. Deployments are now continuous and invisible.
>1MData-virtualisation viewsViews created on the data virtualisation layer, operated around the clock with full lineage.

Context

The institution's daily operations rest on several critical systems: transaction processing at scale, regulatory reporting, and cross-border financial messaging in real time. We have worked on that estate since 2007, and the remit has grown with the institution's own technology transformation, from the fundamental plumbing to the governance decisions.

The engagement now covers the full platform lifecycle, including complete ownership of the data virtualisation layer, which was deployed without any pre-existing internal expertise and is operated by us end to end.

What made it hard

  • A fragile point-to-point legacy estateReplacing a web of point-to-point integrations with a scalable ESB, without disturbing production systems processing millions of transactions a day. Every migration had to be surgical and invisible.
  • Zero downtime, in a 24/7 institutionCritical operations never stop. Every deployment, migration and update had to be invisible to end users and downstream systems alike.
  • Volume and reliability at scale8 million API calls a day with zero errors, continuous real-time monitoring and automatic failover, while business teams expected near-real-time access to data that previously took up to an hour to arrive.
  • Regulatory alignment on a fixed calendarContinuous alignment with SWIFT ISO 20022, deployed globally and in synchronisation across every region each November, under strict traceability and security requirements.

What we did

Infrastructure modernisation and ESB migration

Migration from the legacy architecture to a modern enterprise service bus. Mirror servers made daytime updates possible with zero interruption, which permanently removed the weekly maintenance windows. IBM Integration Bus migrated to IBM App Connect Enterprise, and WebSphere MQ to Apache Kafka for real-time event streaming.

  • IBM IIB → ACE
  • WebSphere MQ → Kafka
  • ESB
  • Mirror servers
  • Zero downtime
Platform diversification and governance

A suite of specialised platforms deployed and governed: an API management layer carrying 8 million calls a day, centralised authentication across every platform in place of the legacy federation service, and a shared environment for the data science teams.

  • API management
  • Centralised SSO
  • Data science platform
Data virtualisation, under full ownership

A virtual abstraction layer that lets teams reach data at source without moving it, and a single point of passage for every reporting tool, which is what guarantees one version of the truth. More than a million views created, with complete data lineage, per-profile security and automated DevOps pipelines.

  • Data virtualisation
  • >1M views
  • Data lineage
  • High availability
Optimisation, security and long-term governance

The role shifted from builder to long-term custodian: rationalising integration flows, cloud migration, security hardening, on-call support and continuous maintenance, with a real-time dashboard for data quality monitoring.

  • Cloud migration
  • Security hardening
  • On-call
  • SWIFT ISO 20022

What comes next

The institution is moving progressively towards cloud-native integration platforms, while the most complex flows stay on the current stack for several more years. The data virtualisation platform keeps growing with new sources and new consuming teams.

Technical stack

  • IBM App Connect Enterprise (ACE)
  • IBM Integration Bus (IIB)
  • WebSphere MQ
  • Apache Kafka
  • SWIFT MT → MX / ISO 20022
  • API management
  • Data virtualisation
  • Real-time monitoring dashboards

The offers behind this engagement

Asset and wealth management

The SWIFT hub every financial flow passes through, in and out

Equities, bonds and derivatives generate a chain of SWIFT messages for every order: instruction, acknowledgement, agent confirmation, settlement. The team that transforms, routes and supports all of them is ours.

An international multi-asset asset management company · Partnership since 2006 · one of our first clients

Asset and wealth management
1000+SWIFT messages a dayEvery trading order generates several round trips, all handled by the team.
200Active ITX mappingsPure IBM TX mappings maintained and evolved, covering every inbound and outbound flow.
~50Automation scriptsShell, KSH and Bash scripts for real-time queue monitoring, delay detection and alerting. No manual watch.
20+ yrsContinuous partnershipPresent since 2006, from design through to production support.

Context

This asset manager runs multi-asset portfolios on international financial markets and handles time-critical transaction flows. Its SWIFT team is composed entirely of our consultants and is the entry and exit point for every financial flow in the business: everything that comes in and goes out passes through it.

The client has functional analysts in house but no technical analysts, so the team naturally extended its remit to that role as well, translating between the business — finance, trading, middle office — and what the integration systems actually do.

What made it hard

  • Market cut-offs, and zero toleranceMarket hours are not negotiable. At precise points in the day, orders must be transmitted. An incident still open at the cut-off has direct financial consequences, so support has to be immediate and final.
  • Two roles in one teamWith no technical analysts on the client side, the team bridges business need and technical implementation as well as building and running the flows.
  • The annual SWIFT November releaseEvery November brings format updates tied to new taxes, regulations and international standards. Every affected message has to be re-analysed, updated and retested. Some years that is transparent; some years it is a project.
  • Around 600 conversion mappingsThe mapping estate covers routing, filtering and transformation. Every version migration means recompiling and retesting the whole set, which is a project in its own right each time.

What we did

Transformation and routing of financial flows

Bidirectional transformation of every flow: internal formats converted into SWIFT messages and back again. Each trading order generates several messages that all transit through the team, alongside routing, filtering and secure archiving over queues and SFTP.

  • IBM TX / ITX
  • SWIFT MT
  • CSV → XML
  • IBM MQ
  • SFTP
  • Secure archiving
Critical operational support and middle office

Support is the largest part of the daily work. The team acts directly on urgent financial incidents on middle-office instruction: cancelling orders, correcting a tax or an amount in flight, issuing corrected messages. When it is too late in the workflow to act from the source application, the team acts on the flow itself.

  • Level 2/3 support
  • Cut-offs
  • Middle office
  • SWIFT correction
Automation, alerting and version management

Around 50 Shell, KSH and Bash scripts on Unix and Linux automate operational surveillance: alerts on blocked queues, detection of late files, incident notifications. Alongside that, the team runs the transformation-engine version migrations and the annual SWIFT release.

  • Shell / KSH / Bash
  • Unix / Linux
  • Automated alerting
  • Version migration

Technical stack

  • IBM TX / ITX
  • SWIFT MT
  • IBM MQ
  • SFTP
  • CSV / XML
  • Shell, KSH, Bash on Unix and Linux
  • Portfolio management system interfaces

The offers behind this engagement

Your flows, mapped the same way.

A 5 to 15 day assessment gives you a flow map, a gap analysis and a costed plan. Fixed price, written to be signed by a decision-maker.

Book an assessment