LSN Logging Buffer Memory Allocation for Network Translation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large scale network address translation (LSN) generates enormous amounts of data, leading to performance slowdowns as the number of sessions increases, due to the inability of traditional logging systems to efficiently manage and process this high volume of information.
Innovation Solution
A system and method where a first device, acting as an intermediary between clients and servers, allocates memory for packet engines, logs LSN information in a logging buffer, and transmits aggregated data to a logging server, while optimizing memory usage by skipping redundant log messages for certain sessions and synchronizing session information across devices using hash keys and transport layer connections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional logging systems are used for LSN sessions, then logging functionality is provided, but system performance slows down as the number of sessions increases
Solution Approach 1:
The patent divides the logging system into multiple components: packet engines that generate logs, a logging buffer that temporarily stores log data, and a logging server that processes aggregated logs. This segmentation allows logging operations to be distributed and asynchronous, preventing performance degradation during high-volume LSN sessions
Solution Approach 2:
The patent introduces a logging buffer as an intermediary component between packet engines and the logging server. This buffer absorbs the shock of high-volume log generation by temporarily storing log data in allocated memory portions, allowing packet engines to continue processing LSN sessions without being blocked by logging operations
2Loss of information
If memory is allocated for logging each LSN session, then complete logging information is recorded, but memory consumption increases with session volume
Solution Approach 1:
The patent allocates specific portions of memory to the logging buffer for each packet engine based on its needs and capacity. This localized memory allocation ensures that each packet engine has dedicated storage for its log data without requiring the entire system memory to be reserved for all possible sessions, optimizing memory utilization
Solution Approach 2:
The patent implements log message skipping for certain sessions based on predefined criteria. Instead of logging every single LSN session, the system selectively logs only necessary information or samples from high-volume sessions, reducing memory consumption while maintaining sufficient monitoring capability
3Loss of information
If log messages are stored for every LSN session, then detailed session information is captured, but redundant data increases processing overhead
Solution Approach 1:
The patent merges log data from multiple packet engines into a single logging buffer, and then aggregates this data before transmitting to the logging server. This consolidation eliminates redundant log entries for sessions that span multiple packet engines and reduces the total volume of data that needs to be processed and stored
Solution Approach 2:
The patent changes the state of log messages by implementing skipping logic that filters out redundant entries based on session parameters. This parameter-based filtering transforms the logging process from recording every event to recording only significant changes or unique sessions, reducing processing overhead
4Stability of the object's composition
If synchronization is implemented across multiple devices, then session information consistency is maintained, but communication overhead increases
Solution Approach 1:
The patent implements periodic synchronization between packet engines and the logging server, rather than continuous real-time synchronization. Log data is accumulated in the buffer and synchronized at intervals or when thresholds are reached, maintaining session information consistency while minimizing communication overhead between devices
Data Source
AI summary
The present invention is related to a method for high volume logging for large scale network address translation. A first device intermediary to a plurality of clients and a plurality of database servers allocates a portion of memory to each packet engine in a plurality of packet engines executing on a respective core of a plurality of cores of the first device. The first device establishes large scale network address translation (LSN) for the plurality of clients, the first device logging LSN information of sessions to a corresponding logging buffer established in a respective packet engine's portion of memory. The first device identifies, for a LSN session, a packet engine from the plurality of packet engines to log the information for the LSN session and stores information of the LSN session to the logging buffer in the packet engine's portion of memory.