VM Memory Restoration Template for Warm Start Latency

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Cloud service providers face inefficiencies in memory management due to idle VMs consuming memory, leading to high costs and limited VM instances per physical machine, exacerbated by high latency in decompressing VM pages during warm starts, which affects performance and resource utilization.

Innovation Solution

Implementing a memory restoration template that predicts and prefetcheds a subset of memory pages from a compressed memory pool to uncompressed memory, reducing software overhead and latency during warm starts, and using hardware accelerators for compression/decompression.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If VM pages are placed in compressed memory pool (ZSWAP) to boost memory capacity, then memory utilization increases and more VMs can be spawned, but decompression latency during warm start increases due to software overhead

Engineering Contradiction:

Improvememory capacityVSAvoiddecompression latency

Core Design Contradiction:

Quantity of substanceVSLoss of time

Solution Approach 1:

The patent applies preliminary action by pre-decompressing memory pages from the compressed memory pool (ZSWAP) before they are actually needed during VM warm start. The system identifies pages that are likely to be accessed during function execution and decompresses them in advance, thereby avoiding the software overhead latency that would occur during on-demand decompression at page fault time.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent segments the memory management process by separating the decompression operation from the page fault handling path. Instead of decompressing pages individually when accessed (on-demand), the system decompresses pages in batches or groups in advance, dividing the workload into manageable segments that can be processed independently before execution.

Inventive Principle:

Principle #1Segmentation

2Productivity

If on-demand decompression is used during page fault, then memory is freed for other uses, but software overhead (lookup in ZSWAP, finding free page) increases latency

Engineering Contradiction:

Improvememory efficiencyVSAvoidpage fault latency

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system performs preliminary decompression of memory pages before they are needed during page fault. By identifying and decompressing pages in advance based on access patterns or execution requirements, the system eliminates the time-consuming lookup and decompression operations that would otherwise occur during critical page fault paths.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent applies the skipping principle by bypassing the traditional page fault handling sequence (lookup in ZSWAP, allocate free page, decompress, copy back). Instead, pages are pre-decompressed and positioned in advance, allowing the system to skip or rush through the critical path operations during actual execution, significantly reducing latency.

Inventive Principle:

Principle #21Skipping (Rushing through)

Data Source

PatentUS20230103447A1Apparatus, Device, Method, and Computer Program for Managing Memory

Publication Date: 2023.04.06 INTEL CORP

AI summary

Examples relate to an apparatus, device, method, and computer program for managing memory, and to a computer system comprising such an apparatus or device. The method comprises determining an impending access to a functionality provided by a virtual machine. The method comprises restoring a subset of memory pages associated with the virtual machine from a compressed memory pool to uncompressed memory according to a memory restoration template after determining the impending access to the virtual machine.