Sliding Hidden Markov Model for Evaluating Discrete Data

Chis, Tiberiu S.

doi:10.1007/978-3-642-40725-3_19

Cited by 6 publications

(3 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Discretized MMPPs (or hidden Markov models) replicate the burstiness of TCP packet traces, which can be clustered in groups, and, hence, allow model parameters to converge on multiple traces simultaneously at reduced computational complexity [10]. Further, arrival parameters of queueing models can be updated incrementally via online EM learning algorithms [2,6,14], which are suitable for live systems.…”

Section: Resultsmentioning

confidence: 99%

Higher response time moments for M/M/1 discriminatory processor sharing queues

Chis¹,

Harrison²

2016

Proceedings of the 9th EAI International Conference on Performance Evaluation Methodologies and Tools

Self Cite

View full text Add to dashboard Cite

Obtaining response time moments in processor sharing (PS) queues is difficult due to serving of multiple jobs. Egaliatarian PS (EPS) queues are limited to one class of arriving jobs. Discriminatory PS (DPS) assigns weights to different job classes and offers more diverse modeling capabilities than EPS. It is known that response time is the representative metric for delay as specified in service level agreements (SLAs), which consider higher moments important. Hence, we build an automated numerical algorithm for calculating higher moments of response time in M/M/1-DPS queues for multiple job classes and test two different case studies.

show abstract

Section: Resultsmentioning

confidence: 99%

Higher response time moments for M/M/1 discriminatory processor sharing queues

Chis¹,

Harrison²

2016

Proceedings of the 9th EAI International Conference on Performance Evaluation Methodologies and Tools

Self Cite

View full text Add to dashboard Cite

show abstract

“…Lube is a geo-distributed framework that reduces the query response times by detecting bottlenecks at runtime [42]. Lube monitors the performance metrics (CPU, memory, network and disk) in real-time and uses Autoregressive Integrated Moving Average (ARIMA) [59] or the Sliding Hidden Markov Model (SlidHMM) [60] to detect resource bottleneck at runtime. The scheduling algorithm considers data locality and bottleneck severity to assign tasks to worker nodes, the late-binding algorithm in Sparrow [61] is used to avoid false positives when detecting bottlenecks by holding a task for a short time before submitting it to a worker node.…”

Section: Spark-based Frameworkmentioning

confidence: 99%

A survey on bandwidth-aware geo-distributed frameworks for big-data analytics

2021

View full text Add to dashboard Cite

In the era of global-scale services, organisations produce huge volumes of data, often distributed across multiple data centres, separated by vast geographical distances. While cluster computing applications, such as MapReduce and Spark, have been widely deployed in data centres to support commercial applications and scientific research, they are not designed for running jobs across geo-distributed data centres. The necessity to utilise such infrastructure introduces new challenges in the data analytics process due to bandwidth limitations of the inter-data-centre communication. In this article, we discuss challenges and survey the latest geo-distributed big-data analytics frameworks and schedulers (based on MapReduce and Spark) with WAN-bandwidth awareness.

show abstract

“…The model is trained on observations until parameter convergence (A, B, ζ, χ, π become fixed). Since the HHMM parameters are trained during runtime on continuously updated response times of components, we use the sliding window technique [28] to read the observations with window size = 10-100 observations, which generates multiple windows with different observations. The obtained observations are used to feed the HHMM.…”

mentioning

confidence: 99%

Detecting and Localizing Anomalies in Container Clusters Using Markov Models

Samir

Pahl

2020

Electronics

View full text Add to dashboard Cite

Detecting the location of performance anomalies in complex distributed systems is critical to ensuring the effective operation of a system, in particular, if short-lived container deployments are considered, adding challenges to anomaly detection and localization. In this paper, we present a framework for monitoring, detecting and localizing performance anomalies for container-based clusters using the hierarchical hidden Markov model (HHMM). The model aims at detecting and localizing the root cause of anomalies at runtime in order to maximize the system availability and performance. The model detects response time variations in containers and their hosting cluster nodes based on their resource utilization and tracks the root causes of variations. To evaluate the proposed framework, experiments were conducted for container orchestration, with different performance metrics being used. The results show that HHMMs are able to accurately detect and localize performance anomalies in a timely fashion.

show abstract

Sliding Hidden Markov Model for Evaluating Discrete Data

Cited by 6 publications

References 9 publications

Higher response time moments for M/M/1 discriminatory processor sharing queues

Higher response time moments for M/M/1 discriminatory processor sharing queues

A survey on bandwidth-aware geo-distributed frameworks for big-data analytics

Detecting and Localizing Anomalies in Container Clusters Using Markov Models

Contact Info

Product

Resources

About