Distributed Peer Appliance Architecture for Scalable Performance Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Legacy performance management systems in computer information systems face scalability issues as they become overloaded with performance data, leading to reduced effectiveness and increased costs when trying to maintain data granularity and processing capacity.
Innovation Solution
A peer-to-peer architecture is implemented, where each peer appliance receives and stores performance data from a subset of computing devices, allowing for distributed data storage and processing, enabling granular performance data and scalability by coordinating data retrieval and processing across multiple appliances.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a single central reporting computer receives data from many collectors, then data collection is simplified, but the reporting computer becomes overloaded and scalability is limited
Solution Approach 1:
The patent divides the centralized reporting computer into multiple distributed peer appliances, each handling a subset of computing devices. This segmentation allows the system to scale horizontally by adding more peer appliances without overloading a single central computer, while each peer maintains simplified data collection for its assigned subset of devices.
Solution Approach 2:
The patent transitions from a single-dimension centralized architecture to a multi-dimensional distributed peer-to-peer architecture. By adding the dimension of distribution across multiple peers, the system achieves both operational simplicity (through standardized peer interfaces) and scalability (through horizontal expansion capabilities).
2Productivity
If the central reporting computer's processing and storage hardware capability is increased, then more performance data can be processed, but the system becomes more expensive
Solution Approach 1:
Instead of concentrating all processing and storage requirements in a single expensive central computer, the patent segments these requirements across multiple peer appliances. Each peer handles a portion of the data processing and storage, allowing the system to achieve high overall capacity through distributed, cost-effective hardware rather than one expensive centralized system.
Solution Approach 2:
The patent combines the processing and storage capabilities of multiple peer appliances to achieve the equivalent or superior capacity of a single large central computer, while distributing the cost across multiple standard hardware units. This merging of distributed resources provides the same productivity at lower individual hardware costs.
3Productivity
If the granularity of performance data is reduced, then there is less data to process, but the system becomes less useful for diagnosing short-duration issues
Solution Approach 1:
The patent implements dynamic data collection where each peer appliance collects granular performance data at appropriate intervals for its assigned computing devices. The system adaptively manages data granularity based on the specific monitoring needs and allows retrieval of detailed granular data when required for diagnosing short-duration issues, rather than using a fixed coarse granularity for all scenarios.
Solution Approach 2:
Each peer appliance maintains high-granularity performance data for its local subset of computing devices, ensuring measurement precision is preserved where needed. The system allows different granularity levels to coexist, with detailed local data retained at each peer for diagnostic purposes while enabling efficient processing through selective data aggregation and retrieval.
4Productivity
If a peer-to-peer architecture is implemented with distributed storage, then scalability and data granularity are improved, but system complexity increases
Solution Approach 1:
The patent makes each peer appliance universal by designing it to perform multiple functions: collecting performance data from computing devices, storing data locally, processing queries, and collaborating with other peers. This multi-functionality simplifies the overall architecture by eliminating the need for specialized components, as each peer can handle various tasks independently while contributing to the distributed system's scalability.
Solution Approach 2:
Each peer appliance is designed to be self-sufficient, maintaining its own local datastore and processing capabilities. Peers independently manage their assigned computing devices' data collection and can autonomously respond to queries using local data, reducing the coordination overhead and complexity compared to a highly centralized system requiring extensive inter-component communication.
Data Source
AI summary
A performance management system, having a peer-to-peer architecture, enables performance transparency in computer information systems, providing granular performance data and scalability. Peer appliances in a computer information system collect performance data. When a user requests a performance report, an originating peer appliance may determine which peer appliances contain the data required for the report and what data processing, if any, is required. The originating peer appliance may send requests indicating what data and what data manipulation processing is required. Each of the receiving peer appliances (including the originating peer appliance) may perform its own portion of the data processing. The originating peer appliance may receive resultant data from the peer appliances (including itself) and combine the resultant data into the requested report for the user. The performance management system distributes significant data processing across the peer appliances, avoids bottlenecks, and increases system scalability.