Multi-Cluster Catalog Service for Hybrid Cloud Replication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional virtual computing systems face challenges in efficiently managing and replicating catalog items across multiple clusters, particularly due to high latency and bandwidth costs when copying images between clusters, and difficulties in moving workloads between on-premises and cloud deployments.
Innovation Solution
A multi-cluster catalog service that manages images and entities by selecting clusters based on available capacity, allowing for minimal latency and bandwidth usage, and integrates with an object store for storing catalog items, enabling policy-based or user-based replication and hybrid cloud functionality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If catalog items are copied between clusters using conventional methods, then the catalog items can be replicated across clusters, but high latency and bandwidth costs occur
Solution Approach 1:
The system performs preliminary actions by pre-fetching and caching catalog item data in local storage before it is actually needed for deployment. The catalog service proactively manages data replication and caching, so when a deployment request occurs, the data is already available locally, eliminating the latency of real-time network transfers.
Solution Approach 2:
The patent introduces a catalog service as an intermediary layer between the network storage and the deployment system. This catalog service maintains local caches and manages data replication, acting as a mediator that reduces direct network traffic between clusters and minimizes latency for catalog item access.
2Reliability
If catalog items are copied between clusters using conventional methods, then the catalog items can be replicated across clusters, but high bandwidth costs occur
Solution Approach 1:
The system implements local quality by maintaining local copies and caches of catalog item data within each cluster rather than relying on centralized remote storage. Each cluster has its own catalog service that can serve data locally, reducing the need for continuous network bandwidth consumption for data access and replication.
Solution Approach 2:
The patent employs copying strategies where catalog items are replicated to local storage in advance and cached locally. Instead of repeatedly transferring the same data over the network, the system creates local copies that can be served without additional bandwidth consumption, significantly reducing network traffic and associated costs.
3Adaptability or versatility
If workloads are moved between on-premises and cloud deployments, then flexibility and scalability are improved, but system complexity increases
Solution Approach 1:
The catalog service is designed with universality to work across both on-premises and cloud environments through a unified interface. The same catalog service architecture and data structures are used regardless of whether the underlying storage is local or remote, allowing workload migration without requiring different systems or increasing operational complexity.
Solution Approach 2:
The catalog service acts as an intermediary abstraction layer that shields users from the complexity of underlying storage differences. Whether data is stored on-premises or in the cloud, the catalog service provides consistent data access and management interfaces, allowing flexible workload migration while maintaining simple, unified system operations.
Data Source
AI summary
An illustrative embodiment disclosed herein is an apparatus including a processor having programmed instructions to receive, from a client, a request to create a catalog item corresponding to a file, determine an identifier of a cluster manager associated with a cluster having available capacity, send, to the cluster manager, a source location of the catalog item, and cause the cluster manager to store the catalog item from the source location in the cluster.