File Server Placement in Stretched Cluster Environments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cross-site workflows in stretched cluster environments for file services lead to performance inefficiencies due to the need for data traversal between different geographic locations, causing delays and performance issues.
Innovation Solution
Implementing site locality techniques by placing file servers within the same site based on host workload and affinity settings, with failover capabilities to ensure continuous service and using dynamic domain name system (DDNS) for Layer 3 networking to manage IP address compatibility during failovers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If file services are accessed in a stretched cluster environment with hosts at different geographic locations, then the system provides distributed storage capability and high availability, but the data workflow traverses between sites causing performance inefficiency and delays
Solution Approach 1:
The patent applies local quality by placing file servers at the same geographic site as the clients that need to access them. This creates location-specific file server instances rather than using a uniform cross-site access model. The system determines the site location of clients and places file servers locally at those sites, making the service quality adaptive to the client's geographic location rather than applying a single access pattern to all clients.
2Productivity
If file servers are placed locally at the same site as clients, then performance efficiency is improved by reducing cross-site data traversal, but system complexity increases due to need for placement algorithms and failover mechanisms
Solution Approach 1:
The patent applies dynamics by implementing dynamic file server placement that adapts to changing system conditions. The placement algorithm continuously monitors client site locations and determines optimal file server placements based on current conditions. When clients move between sites or new clients are added, the system dynamically adjusts file server placement to maintain optimal performance, rather than using a static placement configuration.
Solution Approach 2:
The patent applies preliminary action by proactively placing file servers at appropriate sites before clients need to access them. The system uses placement algorithms to determine optimal file server locations in advance, and implements failover mechanisms that pre-configure backup file servers at alternative sites. This preliminary preparation ensures that when clients need file services, locally-placed file servers are already available, avoiding the complexity of real-time migration during active operations.
3Reliability
If file servers are migrated between sites for failover, then service continuity is maintained, but IP address compatibility issues arise in Layer 3 networking environments
Solution Approach 1:
The patent applies the intermediary principle by introducing a site-aware file server placement system that acts as a mediator between clients and file servers. This placement algorithm serves as an intermediary layer that resolves the conflict between failover requirements and IP address compatibility. It determines which site should host the file server based on client location and network configuration, preventing IP address conflicts that would otherwise occur during cross-site migrations in Layer 3 networking environments.
Data Source
AI summary
The location of resources for file services are located within the same site, thereby eliminating or reducing performance issues caused by cross-site accesses in a stretched cluster environment. A file server placement algorithm initially places file servers at a site based at least in part on host workload and affinity settings, and can perform failover to move the file servers to a different location (e.g., to a different host on the same site or to another site) in the event of a failure of the host where the file servers were initially placed. File servers may be co-located with clients at a location based on client latencies and site workload. Failover support is also provided in the event that the sites in the stretched cluster have different subnet addresses.