Distributed Backup Storage Using Server Advertisement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional data backup methods involve lengthy restore times due to the use of low-performance storage devices, leading to inconvenience and inefficiency, especially when monolithic storage arrays are costly and add complexity to data centers.
Innovation Solution
A distributed backup method utilizing a server group with identical backup and storage policies, where each server transmits advertisement messages for available storage space and selects a destination for backups within the group using a high-speed network, storing backups on high-speed devices initially and moving them to low-speed devices after a certain age.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If backups are stored on low-performance storage devices, then storage capacity is increased, but restore times become lengthy
Solution Approach 1:
The backup system segments storage resources by creating a distributed network of storage nodes within the server group. Each server can act as both a compute node and a storage node, dividing the backup storage function across multiple devices rather than relying on a single monolithic storage array. This segmentation allows recent backups to be stored on high-speed devices for quick restoration while maintaining overall storage capacity through the distributed network.
Solution Approach 2:
The system dynamically assigns backup destinations based on current storage availability and performance characteristics. The distributed backup method allows flexibility in selecting which server within the group will receive backups, enabling optimization of restore performance by directing recent backups to high-speed storage devices while maintaining capacity through utilization of all available storage resources in the group.
2Loss of time
If monolithic consolidated storage arrays are used, then restore times are improved, but device complexity and cost increase
Solution Approach 1:
The invention merges the storage function with the existing server group infrastructure. Instead of adding a separate monolithic storage array with its own switching and cabling complexity, the system combines backup storage capabilities with the servers already present in the data center. Each server in the group can provide storage resources, eliminating the need for dedicated storage hardware and reducing overall system complexity.
Solution Approach 2:
The servers in the group serve multiple functions: they execute services, provide compute resources, and simultaneously act as backup storage devices. This multi-functionality eliminates the need for specialized monolithic storage arrays, as the same servers that perform computational tasks also provide backup storage capabilities, thereby reducing device complexity and cost while maintaining restore performance.
3Loss of time
If distributed backup across server group is implemented, then restore times are reduced, but network bandwidth consumption increases
Solution Approach 1:
The system performs preliminary actions by maintaining backup copies distributed across the server group in advance. When restoration is needed, the system can immediately retrieve data from the distributed backup sources without needing to transfer large amounts of data across the network at restore time. The advertisement messages pre-establish knowledge of available backup locations, enabling efficient restoration by reducing actual data transfer requirements during critical restore operations.
Data Source
AI summary
A distributed backup method includes defining a server group having a plurality of server computers, and providing each server computer with an identical backup policy and an identical storage policy. The method also includes transmitting, at each server computer of the server group to every other server computer of the server group, advertisement messages describing available storage space for storing backups, and receiving, at each server computer from the server, at least some of the advertisement messages. The method also includes identifying, at a first server computer using the backup policy, at least a first service to be backed up, selecting a second server computer using the received advertisement messages, transmitting, from the first server computer to the second server computer, a backup including information corresponding to the at least one service, and storing the backup, by the second server computer using the storage policy.