Log-less IaaS Snapshotting for Cloud Workload Protection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing systems for protecting IaaS workloads in a public cloud environment are inefficient, particularly in recreating virtual machine data across datacenters in case of failures, often relying on IO log files that consume computing resources and are inefficient.

Innovation Solution

Implementing a log-less, guest-agnostic snapshotting system that uses native hypervisor features like Hyper-V Replica and multi-stage snapshot infrastructure to create and replicate virtual hard drive snapshots directly, eliminating the need for virtual machine agents and IO log files, and utilizing Field-Programmable Gate Arrays for fast interconnects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If an agent executing in the virtual machine keeps IO log files to enable data recreation, then failure protection is improved, but computing resources are reduced and efficiency deteriorates

Engineering Contradiction:
Improvefailure protectionVSAvoidcomputing resource efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent extracts the logging function from the virtual machine agent and relocates it to the host system. The host keeps IO log files instead of the guest agent, separating the logging burden from the virtual machine's computing resources while maintaining failure protection capabilities

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent introduces an intermediary layer (the host system acting as a mediator between storage and computation) that manages IO logging independently. This intermediary approach allows the virtual machine to focus on computation while the host handles logging, resolving the resource conflict

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If virtual machine agents are used to maintain IO log files, then data protection is improved, but device complexity increases

Engineering Contradiction:
Improvedata protectionVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent removes the logging functionality from the virtual machine agent, eliminating the complexity of maintaining logging agents within each virtual machine. The logging function is extracted and consolidated at the host level, simplifying the overall system architecture

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The host system assumes a universal role, performing both virtual machine management and IO logging functions. This consolidation reduces the number of separate components needed, as the host becomes a multi-functional platform that eliminates the need for dedicated logging agents in each VM

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Reliability

If IO log files are maintained by virtual machine agents, then failure recovery is enabled, but recovery time increases

Engineering Contradiction:
Improvefailure recovery capabilityVSAvoidrecovery time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent implements preliminary action by having the host continuously maintain IO log files in advance of any failure event. This pre-positioned logging infrastructure eliminates the need for agents to start logging at the time of failure, enabling immediate recovery operations without agent initialization delays

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3750066B1Protection of infrastructure-as-a-service workloads in public cloud
Publication Date: 2023.05.10 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP3750066B1 patent drawingFigure 1
  • EP3750066B1 patent drawingFigure 2
  • EP3750066B1 patent drawingFigure 3

AI summary

A system may include a cloud-based virtual machine connected to multiple virtual hard drives and associated with a guest of a first datacenter. A recovery management server may, substantially simultaneously for each virtual hard drive, create an Infrastructure-as-a-Service snapshot of data on the virtual hard drive without utilizing an agent of the virtual machine. The recovery management server may also directly replicate the snapshot of data for each virtual hard drive at a second datacenter, which may be geographically remote from the first datacenter, without re-creating Input Output ("IO") operations for each virtual hard drive via a log file. Upon an indication of failure of the first datacenter, the recovery management server may arrange to have each virtual hard drive's replicated snapshot of data be consumed by a recovery service for the guest at the second datacenter.