Base Caching Disk for Fast Virtual Machine Instantiation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The instantiation of virtual machines in cloud computing systems is hindered by significant network traffic congestion and latency due to the need for multiple duplicate copies of images to be retrieved from a central image store, leading to delays of up to twenty-one minutes, which negatively impacts user experience.
Innovation Solution
Implementing a base caching disk coupled with differencing disks that perform on-demand fetching of data sections, allowing only one copy of the image to be downloaded from the central image store, and caching subsequent requests on the base caching disk to reduce latency and load on the central image store.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple duplicate copies of images are retrieved from a central image store for virtual machine instantiation, then complete virtual machine booting is achieved, but network traffic congestion and instantiation latency increase significantly
Solution Approach 1:
The patent applies preliminary action by pre-fetching and caching image data sections to the base caching disk before virtual machine instantiation is fully completed. The system retrieves and stores image sections in advance, so that when instantiation occurs, the data is already available locally, eliminating the need to wait for complete image retrieval during booting.
Solution Approach 2:
The patent segments the complete virtual machine image into multiple separate data sections that can be independently retrieved and cached. Instead of downloading the entire image as a single unit, the system divides it into manageable sections that can be fetched separately and stored in the base caching disk, allowing partial instantiation to proceed while remaining sections are being retrieved.
2Reliability
If multiple duplicate copies of images are downloaded from the central image store, then each virtual machine receives its complete image, but the number of retrieval requests increases network load
Solution Approach 1:
The patent merges multiple retrieval requests into a single caching operation. Instead of each virtual machine instantiation triggering separate downloads from the central image store, the system consolidates image section retrieval into the base caching disk, which then serves multiple virtual machines from the cached copies, reducing the number of requests to the central store.
Solution Approach 2:
The base caching disk acts as an intermediary between the central image store and the virtual machines. Rather than virtual machines directly requesting images from the central store (creating multiple duplicate retrieval operations), the base caching disk intercepts these requests, serves cached copies when available, and only contacts the central store for sections that need to be pre-fetched.
3Reliability
If the entire image is downloaded before virtual machine booting, then complete image retrieval is ensured, but instantiation time increases to five to twenty-one minutes
Solution Approach 1:
The patent applies partial action by retrieving and caching only the necessary image sections required for virtual machine instantiation rather than downloading the entire complete image in advance. The base caching disk stores specific data sections that are needed for booting, allowing instantiation to proceed with partial image availability while remaining sections are fetched as needed.
4Reliability
If duplicate image copies are retrieved for each virtual machine, then each machine has its own image copy, but network traffic congestion occurs
Solution Approach 1:
The patent uses copying by creating image section copies in the base caching disk that can be served to multiple virtual machines. Instead of retrieving duplicate copies from the central image store for each virtual machine (causing network congestion), the system copies image sections once to the base caching disk, which then provides local copies to multiple virtual machine instantiation requests.
Data Source
AI summary
A base caching disk and associated methods for facilitating fast instantiation of virtual machines in distributed computing systems are disclosed herein. In one embodiment, a method includes during an instantiation process of a virtual machine on a node in the distributed computing system, determining whether a base caching disk on the node contains a block of data of the image needed for the instantiation process of the virtual machine. The method can also include, in response to determining that the base caching disk does not contain the block of data, retrieving the block of data from the central image store without retrieving the entire image, writing the retrieved block of data to a sector on the base caching disk at the node and subsequently, marking the sector containing the written block of data on the base caching disk as read-only.