Scaling XTRF To Semantix Requirements and System Integrations

A successful joint effort to scale up from single server node to multi-node with one of the largest LSPs in the world

Semantix, the leading language company in the Nordic region, is geared for expansion and requires a system flexible enough to expand with their needs, and with the capability of running multiple nodes (servers) for a variety of translation tasks, across many clients. XTRF worked closely with Semantix to adapt the system to these requirements. First, some background on Semantix.

Semantix was founded in 1969 and is now the thirteenth largest language service provider in the world with over 400 employees, 13 offices in 7 countries, and working in a staggering range of language pairs, over 238 and counting. With an average 80,000 projects annually, the challenge of consolidating workflows and coordinating those workflows across various systems, was a daunting one. Their strategic focus on language technology and being a dynamic customer partner further raised the bar.

Semantix became an XTRF customer through acquisition. Through evaluating XTRF against other options in place at the Semantix group, they decided to move all operations into XTRF and scale up to 200 users by the end of 2020. The decision at Semantix was largely based on the ability for XTRF to handle complex localization workflow requirements and bespoke customer solutions. This required a plan for scaling the systems to handle an extremely high order flow while maintaining connections to other tools in their translation workflow stack, including CAT, Accounting and Invoicing, CRM, and more. In addition, they required a modular approach that would allow them to add nodes (servers) and use a custom load balancer application to rout jobs and any associated metadata to the properly configured node.

A true technology partnership

During a twelve month implementation program, engineering teams from Semantix and XTRF worked together to build the multi-node system and to configure the required APIs and connectors to connect to other critical systems for billing, customer management, and more. The goal was an automation level that eliminates manual connections between departments, streamlining the customer experience for their translation and interpreting clients.

Semantix presented the results of this collaboration at the XTRF Summit 2019 to showcase the potential for high scalability in a very complex language business. Connecting software resources, the introduction of the load balancing application that coordinates traffic between multiple server nodes, and other integrations were developed to fully automate internal workflows.

Why Semantix chose XTRF

Semantix had a list of requirements, including:

  • A Modular Approach
  • Automation of Processes
  • Workflow Optimization
  • CAT tool independency
  • Scalability

XTRF was able to provide the required scalability, interconnected workflow optimization, and a modular approach, while maintaining the independence of the CAT system. Engineering was required to develop the multi-node scalability required, and both teams worked together to synchronize the required API connections between multiple applications.

These synchronizations included API – Client portal and Home portal, Macros & periodic jobs:

  • Exports invoice data to Accounting system
  • Exports views and reports to relevant users/systems
  • Adjusts workflows
  • Checks for potential issues
  • File archiving
  • Batch updates on misc. Objects

A modular approach and flexible connectivity create a holistic system

A modular approach was adopted to optimize and automate workflows, while offering a wide array of configurations depending on the nature of the work, client requirements, etc. Today, jobs flow directly from client systems into Semantix and to any client-specific vendors, invoices are generated automatically and matched to jobs, and customer information is updated into the CRM across two systems.

Technical Challenge

Multi-node adoption to manage a wide variety of associated factors

The decision to adopt a multi-node model was necessary to handle the wide range of requirements and the high volume of projects at any one given time. Incoming projects from the Internet enter the Load Balancer which parses them and directs each project to the proper XTRF node (server). Each node manages the project and then populates the central database and stores the project in a dedicated file storage location. Any associated metadata is retained and stored for future needs.

XTRF and Semantix developed a deployment plan with timeline of twelve weeks. The deployment plan (12 weeks) defined the following areas for development and coordination:

  • Define successful criteria of the solution
  • Aligning all customizations and integrations
  • Coordinate UAT across departments
  • Continuous development
  • Partnership

Future Refinement

This project is ongoing with many future improvements planned including:

  • Improved user experience
  • Less downtime
  • Better performance
  • Continued open collaboration

A primary consideration is the optimization of the Load Balancer including more control of what goes into each node (separating traffic based on tasks, security needs,  and expected loads.). Fortunately the system design is robust enough to support additional development and scaling for the foreseeable future.

This project demonstrates the scalability and flexibility of the XTRF platform, and our willingness to work with our customers to provide the optimum solutions for their needs. The goal is a win-win scenario for everyone involved.