Enclosure Cache Allocation for Server Workloads

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Modern computing systems face inefficiencies in using non-volatile memory as cache due to high costs and the need for over-provisioning to cover all workloads and hotspot sizes, making it uneconomical to provide sufficient cache for all scenarios.

Innovation Solution

An enclosure cache system that determines the cache hit rate and available cache for each server, dynamically allocating and offering a portion of the enclosure cache based on these factors to optimize cache utilization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If direct attached cache is over-provisioned to cover all possible workloads and hotspot sizes, then cache performance and latency are improved, but cost increases significantly

Engineering Contradiction:

Improvecache performanceVSAvoidcache memory capacity

Core Design Contradiction:

ReliabilityVSQuantity of substance

Solution Approach 1:

The patent implements dynamic cache allocation where the enclosure cache is not statically assigned but dynamically offered to enclosure attached servers based on their workload characteristics, cache hit rates, and available cache resources. This allows the system to adapt cache distribution in real-time rather than over-provisioning for all possible scenarios simultaneously

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes the parameter of cache allocation from fixed/static to variable/dynamic by continuously monitoring cache hit rates and adjusting the amount of enclosure cache offered to each server. This enables the cache capacity parameter to be optimized based on actual workload conditions rather than worst-case scenarios

Inventive Principle:

Principle #35Parameter changes

2Quantity of substance

If non-volatile memory is used for direct attached cache, then cost is reduced compared to DRAM, but performance and access time are worsened

Engineering Contradiction:

Improvecache memory capacityVSAvoidcache access time

Core Design Contradiction:

Quantity of substanceVSSpeed

Solution Approach 1:

The patent merges two different cache types into a unified cache hierarchy: local server cache (typically faster memory) and enclosure cache (non-volatile memory). By combining these caches and presenting them as a unified local cache to the server, the system achieves both the capacity of non-volatile memory and the performance of faster memory through intelligent caching strategies

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The enclosure cache acts as an intermediary between the slow non-volatile storage and the fast local server cache. It captures and caches frequently accessed data from storage, reducing the need for servers to access slow storage directly, while the local server cache handles the most frequently accessed data for even faster access

Inventive Principle:

Principle #24Intermediary (Mediator)

3Productivity

If enclosure cache is dynamically allocated based on cache hit rate and available cache, then resource utilization efficiency is improved, but system complexity increases

Engineering Contradiction:

Improvecache utilization efficiencyVSAvoidcache management system

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The system implements feedback mechanisms where the enclosure monitors cache hit rates from each server and uses this feedback to adjust the amount of enclosure cache offered to each server. This closed-loop control enables automatic optimization of cache distribution without manual intervention, improving utilization while keeping management complexity hidden from users

Inventive Principle:

Principle #23Feedback

4Adaptability or versatility

If cache is allocated to cover all possible workloads, then adaptability to different workloads is improved, but cost increases due to over-provisioning

Engineering Contradiction:

Improveworkload coverageVSAvoidcache memory capacity

Core Design Contradiction:

Adaptability or versatilityVSQuantity of substance

Solution Approach 1:

The enclosure cache serves multiple functions and multiple servers simultaneously, acting as a universal cache resource that can be dynamically allocated to different servers based on their needs. This multi-functional approach allows a single cache resource pool to serve diverse workloads without requiring separate dedicated caches for each server or workload type

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS9176854B2Presenting enclosure cache as local cache in an enclosure attached server

Publication Date: 2015.11.03 LENOVO GLOBAL TECHNOLOGIES SWITZERLAND INTERNATIONAL GMBH

AI summary

Presenting enclosure cache as local cache in an enclosure attached server, including: determining, by the enclosure, a cache hit rate for local server cache in each of a plurality of enclosure attached servers; determining, by the enclosure, an amount of available enclosure cache for use by one or more of the enclosure attached servers; and offering, by the enclosure, some portion of the available enclosure cache to an enclosure attached server in dependence upon the cache hit rate and the amount of available enclosure cache.