Memory Cache Data Center Segmentation for Latency Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

As the number of users increases, managing data across a single data center becomes challenging, and network latency issues arise due to users being located far from the data center, especially when large data centers lack efficient caching mechanisms.

Innovation Solution

Implementing a memory cache data center that communicates with other data centers to store and retrieve data, using a memory cache that does not include a local database, allowing it to focus on volatile storage with proactive backfilling and data replication across multiple servers for redundancy and efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If a single large data center is used to manage all users, then data storage capacity is sufficient, but management complexity increases and network latency increases for distant users

Engineering Contradiction:
Improvedata storage capacityVSAvoidmanagement complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent divides the single large data center into multiple distributed data centers (first data center, second data center, etc.), each managing a portion of the user base. This segmentation reduces management complexity while maintaining sufficient storage capacity across the distributed system, and places users geographically closer to their local data center to reduce network latency.

Inventive Principle:
Principle #1Segmentation

2Quantity of substance

If a single large data center is used to manage all users, then data storage capacity is sufficient, but network latency increases for distant users

Engineering Contradiction:
Improvedata storage capacityVSAvoidnetwork latency
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

By segmenting the user base across multiple geographically distributed data centers, users experience reduced network latency as they connect to their local data center rather than a distant centralized one. The total storage capacity is maintained through the distributed architecture.

Inventive Principle:
Principle #1Segmentation

3Ease of manufacture

If a memory cache without local database is used, then hardware costs are reduced, but data persistence capability is reduced

Engineering Contradiction:
Improvehardware costsVSAvoiddata persistence capability
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The patent introduces a database system as an intermediary between the memory cache and persistent storage. The memory cache provides fast access to frequently used data while the database system handles data persistence and durability, allowing the system to achieve both low hardware costs and reliable data persistence without requiring expensive persistent storage in every node.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP2545458B1Method and memory cache data center
Publication Date: 2018.09.19 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP2545458B1 patent drawingFigure 1
  • EP2545458B1 patent drawingFigure 2
  • EP2545458B1 patent drawingFigure 3

AI summary

A data center system includes a memory cache coupled to a data center controller. The memory cache includes volatile memory and stores data that is persisted in a database in a different data center system that is located remotely from the data center system rather than in the first data center system. The data center controller reads data from the memory cache and writes data to the memory cache.