Host Hypervisor Trace Mining for Virtual Machine Workload Characterization
Bibliographic record
Abstract
The efficient operation and resource management of multi-tenant data centers hosting thousands of services is a demanding task, that requires precise and detailed information regarding the behaviour of each and every virtual machine (VM). Often, coarse measures such as CPU, memory, disk and network usage by VMs are considered in grouping them onto the same physical server, as detailed measures would require access to the guest operating system (OS), which is not feasible in a multi-tenant setting. In this paper, we propose host-level hypervisor tracing as a non-intrusive means to extract useful features, that can provide for fine grain characterization of VM behaviour. In particular, we extract VM blocking periods as well as virtual interrupt injection rates to detect multiple levels of resource intensiveness. In addition, we consider the resource contention rate due to other VMs and the host, along with reasons for exit from non-root to root privileged mode, revealing useful information about the nature of the underlying VM workload. We also use tracing to get information about the rate of process and thread preemption in each VM, extracting process and thread contention as another feature set. We then employ various feature selection strategies and assess the quality of the resulting workload clustering. Notably, we adopt a two-stage feature selection approach in addition to a one shot clustering scheme. Moreover, we consider inter-cluster and intra-cluster similarity metrics, such as the silhouette score, to discover distinct groups of workloads as well as workload groups with significant overlap. This information can be used by 1) data center administrators to gain deeper visibility into the nature of various VMs running on their infrastructure, 2) performance engineers to assist root cause analysis of VM issues and 3) IaaS providers to help in resource management based on VM behavior.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".