Gcore

In an era where gaming experiences demand unparalleled performance and artificial intelligence requires sophisticated computational resources, the right infrastructure can make or break success. Whether you're a indie developer launching a new game or a large studio managing millions of players, the technical foundations supporting your operations are crucial. This article explores how Gcore is revolutionizing the industry through its next-generation infrastructure, combining the demands of gaming with the capabilities of AI in a scalable, secure, and efficient platform.

Gcore's Comprehensive Cloud Services

The company's comprehensive cloud services encompass a wide range of solutions designed to support gaming and AI applications. Gcore's infrastructure, which spans over 180 points of presence (PoPs) and 50 cloud locations, forms the backbone of their service offering.

Infrastructure Services

Gcore provides bare metal servers capable of supporting scalable configurations with up to 128GB of RAM and 4 GPU A2 instances, making them particularly suitable for latency-sensitive gaming applications. Their platform supports DDoS protection for both bare metal and virtual machines, ensuring robust security.

Gaming-Specific Solutions

The company's gaming infrastructure includes virtual machines, bare metal servers, and GPUs that work together to accelerate game builds, testing, and deployment. This setup allows developers to run multiple testing environments on demand, supporting everything from indie studios to major AAA developers handling millions of players.

Performance and Scalability

Gcore's platform automatically scales resources to handle sudden spikes in player activity, critical for maintaining smooth performance during gaming sessions. Their network, built on an architecture that enables L2 connectivity between bare metal and virtual instances, ensures efficient data movement and reduced latency.

CDN and Streaming Services

The company's CDN network features 180 PoPs and 200+ Tbps capacity, designed to accelerate content delivery for gaming and video applications. Integrated transcoding and VOD storage solutions further enhance the platform's capabilities, supporting live events and seasonal game launches.

AI Infrastructure

Gcore's AI infrastructure, built on NVIDIA GPU technology, provides scalable computing resources for AI model building, training, and deployment. This technology supports both AI and high-performance computing workloads, making it adaptable to various gaming-related AI initiatives.

Flexible Bare Metal Solutions

Gcore's bare metal servers deliver minimal lag and uninterrupted gameplay through high-frequency CPUs/vCPUs and dedicated compute power for latency-sensitive applications. The infrastructure supports up to 128GB of RAM and four GPU A2 instances, providing the necessary configuration flexibility for complex gaming workloads.

The platform's pay-as-you-go model with hourly billing enables customers to scale server usage precisely in line with player demand. During peak times, such as major game launches, server capacity can be quickly ramped up to handle increased loads. Conversely, during periods of lower activity, resources can be scaled down to optimize costs without long-term commitments.

For instance, during Saber Interactive's Space Marine 2 launch, the infrastructure successfully managed over 400,000 concurrent players. The system automatically scaled resources to handle the sudden spike in player activity while maintaining consistent performance and preventing server crashes. Similarly, the platform has demonstrated success in supporting genomic research, where GPU Cloud configurations delivered a 95% reduction in processing time while maintaining full flexibility and scalability.

The bare metal servers integrate seamlessly with other Gcore services to create an optimized gaming environment. For example, the infrastructure combines virtual machines, bare metal servers, and GPUs to accelerate game builds, testing, and deployment. The platform also incorporates CDN solutions to reduce download times and minimize patch distribution delays, while managed Kubernetes and load balancers help simplify multiplayer game scaling and prevent server congestion during peak hours. The global network, featuring 180+ points of presence, enables deployment of servers in locations where players are concentrated, further reducing latency and improving the overall gaming experience.

AI-Powered Infrastructure

The company's AI infrastructure built on NVIDIA GPU technology provides scalable computing resources for AI model building, training, and deployment, supporting both AI and high-performance computing workloads.

Performance Optimization

Gcore's solution delivers significant performance improvements through dynamic resource allocation and efficient GPU utilization. Initial benchmark tests demonstrated processing time reductions of up to 50x for variant calling tasks compared to CPU processing, while achieving consistent results across multiple GPU configurations.

Scalable Infrastructure

The infrastructure supports both bare metal GPU servers and virtual GPU instances, allowing customers to choose between direct hardware access for maximum performance or scalable VM options that reduce costs through flexible resource allocation. Gcore's approach enables dynamic scaling based on real-time needs while maintaining data sovereignty through regional processing.

Performance Metrics

Processing times for specific tasks demonstrate the effectiveness of the GPU implementation. For example, the Haplotypecaller task completed in 13 minutes using 2 GPUs, and the Deepvariant task finished in 89.5 minutes with 8 GPUs, with benchmark results consistently showing faster processing times compared to traditional CPU approaches.

Data Sovereignty

The solution maintains data privacy and compliance through regional processing while delivering the required compute performance for high-demand applications. By processing data within the Baltic region, Gcore's infrastructure supports digital sovereignty requirements while providing access to state-of-the-art GPU technology.

Global CDN and Streaming Services

Gcore's CDN network operates on an architecture featuring 180+ points of presence (PoPs) worldwide and 200+ Tbps network capacity. This infrastructure design enables users to access applications as if hosted in their local region, with average latency measured at 30 ms.

The CDN supports both IPv4 and IPv6 protocols and can process up to 2 million requests per second with an 85% average cache hit ratio. The network architecture includes advanced features like Anycast and GeoDNS routing, L2 connectivity between bare metal and virtual instances, and support for HTTP/2 and HTTP/3 (beta) protocols, including TLS 1.3 with 0-RTT connections.

Key Features and Performance

The service offers dynamic content acceleration, improving web application performance and content delivery speed. It includes built-in security features such as advanced access policies, L3, L4, L7 DDoS mitigation, and next-gen Web Application Firewall (WAF) protection. The CDN handles up to 14 Tbps daily peak bandwidth and processes 2 million requests per second.

Use Cases

The CDN has demonstrated significant value across multiple applications. In gaming, it enables content prefetching, optimizes large file delivery, and implements cache sharding to improve latency. For video streaming, it delivers reduced time-to-first-byte (TTFB) for instant playback, maintains stream integrity during large-scale broadcasts, and employs micro-segmented caching for live video delivery.

The infrastructure supports various content optimization features, including image compression and transformation, video transcoding, and custom SSL certificate management. Security features encompass SSL/TLS encryption, web access control, and DDoS protection capabilities. The CDN supports both bare metal and virtual GPU instances, allowing customers to choose between direct hardware access for maximum performance or scalable virtual machine options that optimize costs through flexible resource allocation.

Performance and Scalability

Automated scaling ensures that resources adjust in real-time to fluctuating player activity, preventing server crashes during sudden spikes. The platform demonstrates this capability through its successful management of the Space Marine 2 launch, which saw over 400,000 concurrent players; the system dynamically scaled resources to handle this demand while maintaining consistent performance.

The hybrid infrastructure combines both bare metal and cloud services to efficiently manage player demands while supporting cross-platform multiplayer experiences. The pay-as-you-go model, which allows hourly billing for bare metal servers, enables customers to scale server usage precisely in line with player demand. During the initial launch phase, the system ramped up server capacity to accommodate the influx of players, then scaled down as demand normalized over time.

The infrastructure's built-in DDoS protection keeps gaming servers secure while automatically scaling resources. This comprehensive protection combines L3, L4, and L7 DDoS mitigation with advanced access policies and next-gen Web Application Firewall (WAF) protection, safeguarding both bare metal and virtual machine environments. Together, these features ensure robust security and reliable performance for gaming applications.