At a glance
Solution:
DevOps
Customer:
Global video game development organization
Customer Size:
Corporation (5,000-10,000 employees)
Industry:
Gaming
Location:
France
Background:
A globally renowned video game organization were operating with a large number of Subversion nodes across the globe.
Challenges:
- Globally distributed teams in a large number of locations, difficulty to align projects.
- Coordination challenges: Teams working across multiple time zones were facing significant coordination challenges.
- Coordination challenges: Teams working across multiple time zones were facing significant coordination challenges.
- Slow SCM performance.
- Version control delays: Slow Source Control Management (SCM) resulted in delays when checking in and checking out code, merging branches, and resolving conflicts.
- Non-local interaction delays: Where developers were needing to access SVN data from a source not geographically close, their delays increased exponentially.
- Developer frustration: Developers had become frustrated with sluggish performance.
- Barrier to meeting production deadlines
- Delays in development: Delays in the development process, whether due to SCM performance, coordination challenges, or other factors, were starting to impact production deadlines.
- Dependency management: Managing dependencies between different teams or components was challenging, this led to delays if one part of the project was not ready on time.
- Tool efficiency: The efficiency and effectiveness of development tools directly impact developer productivity.
- Workflow interruptions: Any interruptions due to technical issues or slow systems were found to reduce the time developers spend on actual coding.
- Current home-grown SCM syncing approach is inefficient and prone to issues.
- Lack of scalability: Relying on a home-grown Subversion syncing solution made changes to the node count incredibly complex.
- Reliability concerns: These solutions were not found to be as reliable as the purpose built alternative, leading to more frequent downtime or data sync errors.
- Maintenance burden: Maintaining and updating a custom synced SVN ecosystem was found to be resource intensive, diverting valuable time and resources away from core development activities.
- Maintenance and/or disaster recovery events take too long, can take between 1-2 days.
- Downtime impact: Long maintenance or disaster recovery times had significantly impacted productivity, as developers were unable to access Subversion system during these periods.
- Data loss risk: Extended recovery times increase the risk of data loss or corruption, which can have severe consequences for the project.
Solution:
We installed BlueOptima Subversion MultiSite at 20 sites distributed around the globe.
- This interconnected global infrastructure unified all development teams, enabling the organization to collaborate seamlessly across all geographical locations.
- MultiSite’s selective replication groups allowed administrators to easily position repository replicas as needed.
- This tailor-made repository replication software freed the team from the inefficiencies of custom SVN syncing scripts, allowing them to focus on meeting project deadlines.
- The implementation of BlueOptima MultiSite Plus also enhanced system robustness. During downtime events impacting one site, all other sites remained operational and active, ensuring continuous productivity.
Results:
- The distribution of SVN repositories into ‘replication groups’ allowed the gaming organization admins to ensure replicas of repositories were where they needed to be.
- This provided developers with near-local SVN interaction times for all interactions.
- This expected 3-5% reduction in wait time resulted in an estimated cost saving equal to the average annual cost of 23 full-time software engineers.
Feedback:
The gaming organization commended the functionality of the product in the face of a system down event. Allowing developers to continue to work while the admin team are resolving a high profile issue is highly valuable. The ability to perform routine maintenance to the environment without the need for costly and impactful downtime across all development teams was also a key benefit for the team.