Throttle disk I/O using disk drive simulation model

Abstract

Computerized methods, systems, and computer-storage media for throttling requests from virtual machines (VMs) to a hard-disk drive (HDD) are provided. When a request for disk I / O is received from a VM, a disk-drive model that simulates performance characteristics of the HDD is accessed. During access, the disk-drive model's estimation of HDD parameters and the disk-drive model's estimation of a current state of a disk head of the HDD are gathered. A projected execution time to carry out the request is computed as a function of the estimated HDD parameters and the estimated current state of the disk head. Also, an actual execution time to carry out the request is measured upon allowing the request to pass to the HDD. Using a comparison of the projected execution time and the actual execution time, the traffic of the requests from the VMs is throttled.

Description

technical field [0001] The present invention relates to throttling disk I / O using a disk drive simulation model. Background technique [0002] Large-scale networked systems are common platforms used in a variety of settings for running applications and maintaining data for business and operational functions. For example, a data center (eg, physical cloud computing infrastructure) can simultaneously provide multiple services (eg, web applications, email services, search engine services, etc.) to multiple customers. These massively networked systems typically include a large number of resources distributed throughout the data center, where each resource resembles a physical machine or a virtual machine (VM) running on a physical node or host. When a data center hosts multiple tenants (such as client programs), these resources are optimally allocated from the same data center to different tenants. [0003] Typically, multiple VMs will run concurrently on the same physical nod...

AI Technical Summary

Problems solved by technology

[0005] Accordingly, because the megabytes / sec-based approach (relevant only to sequential access) and the request / sec-based approach (relevant only to random access) each focus on different specific aspects of disk I / O performance, these approaches cannot Consider both sequential and random access to physical disks

For example, megabytes / second based approaches focus on sequential access, thus presenting the risk of being overwhelmed by a flood of random access and generating a backlog of requests

On the other hand, requests / sec based approaches focus on random access, presenting the risk of encountering sequential access strings and unnecessarily limiting throughput to the physical disk

Therefore, the inability to account for sequential versus random access, or the large difference in disk I / O performance they produce, makes these methods inefficient

Images