Kvick, the company behind the open-source Spegel project (Spegel
Kvick, the company behind the open-source Spegel project (Spegel.dev), distributes the production-ready, hardened, and enhanced version of Spegel with additional capabilities. Spegel powers Kubernetes deployments of any size and is used by thousands of enterprises, including the largest AI, financial, streaming and technology companies. 10x Faster deployments No more individual pulls from the registry for each deployment. P2P distribution of cached images across nodes accelerates deployments up to 20x. Benefits increase with increased usage of short-lived containers. Cold Starts Spegel Enterprise, with its tracker, solves the DDOSing issue where each node tries downloading the same image from the source and effectively throttling the network. Reduced egress and idle time costs Up to 90% reduction in egress fees and idle time spent at pod initiation Multi-cluster/cloud distribution Enabling every cluster to be its own stateless registry removes the need for stateful regional registries and the overhead of managing them. AI & Edge AI Workloads Images with AI models exceed tens of gigabytes in size. Increased sizes also make it harder to distribute at the edge. . High Availability Spegel Enterprise users don't get affected by registry reliability issues as every node becomes a distribution node
Kvick, the company behind the open-source Spegel project (Spegel
Kvick, the company behind the open-source Spegel project (Spegel.dev), distributes the production-ready, hardened, and enhanced version of Spegel with additional capabilities. Spegel powers Kubernetes deployments of any size and is used by thousands of enterprises, including the largest AI, financial, streaming and technology companies. 10x Faster deployments No more individual pulls from the registry for each deployment. P2P distribution of cached images across nodes accelerates deployments up to 20x. Benefits increase with increased usage of short-lived containers. Cold Starts Spegel Enterprise, with its tracker, solves the DDOSing issue where each node tries downloading the same image from the source and effectively throttling the network. Reduced egress and idle time costs Up to 90% reduction in egress fees and idle time spent at pod initiation Multi-cluster/cloud distribution Enabling every cluster to be its own stateless registry removes the need for stateful regional registries and the overhead of managing them. AI & Edge AI Workloads Images with AI models exceed tens of gigabytes in size. Increased sizes also make it harder to distribute at the edge. . High Availability Spegel Enterprise users don't get affected by registry reliability issues as every node becomes a distribution node