MI MAURO IACONOIT ↗

COURSE MATERIALS / DATA ANALYTICS

Computer Systems Modeling and Semantic Web

A course on models for computer-based systems, performance evaluation, queueing networks, and stochastic Petri nets, using analytical and simulation methods.

All courses / Computer Systems Modeling and Semantic Web
Versione italiana ↗

A course on models for computer-based systems, performance evaluation, queueing networks, and stochastic Petri nets, using analytical and simulation methods.

The lessons move from performance modelling to probability, queues, stochastic Petri nets, and multi-class systems, connecting model assumptions to useful performance measures.

Slides remain in English, the original language of the supplied materials. Introductory commentary is in English; an Italian version is also available.

University
University of Campania “Luigi Vanvitelli”
Department
Department of Mathematics and Physics
Degree programme
Bachelor’s Degree in Data Analytics (L-41)

Course pathway

The lessons move from performance modelling to probability, queues, stochastic Petri nets, and multi-class systems, connecting model assumptions to useful performance measures.

01 / LECTURES

Theoretical
lectures.

Numbered in the order of the supplied materials.

  1. 01IntroductionThe opening lesson sets the course goals, workload, assessment, and references. It frames modelling as a way to describe computer-based systems and prepare for performance questions developed through queueing and simulation.
  2. 02Introduction to performance models and queueing networksThis lesson introduces systems that receive work, process it, and return results. Queueing networks represent customers, service stations, routing, and congestion, connecting system structure to response time and throughput.
  3. 03Open modelsOpen models represent workloads whose arrivals are generated outside the system. The lesson develops arrival rates, routing, stability, and service demand, preparing students to interpret performance as offered load changes.
  4. 04Analysis of closed modelsClosed models describe a fixed population of users or jobs circulating through a system. The lesson connects population, think time, service demand, throughput, and response time, contrasting closed workloads with open arrivals.
  5. 05Probability and statistics for performance evaluationMeasurements vary because workloads and service times are uncertain. This lesson reviews tools for describing variability, interpreting distributions, and reasoning about estimates rather than relying on a single observation.
  6. 06Probability distributionsThe lesson surveys distributions used to model random quantities in performance studies. It relates assumptions and parameters to the phenomena represented, helping choose a model suited to the system.
  7. 07Fitting and generation of distributionsThis lesson moves from selecting a distribution to estimating parameters from data and generating samples for simulation. It distinguishes fitting a model to observations from generating values according to that model.
  8. 08A closer look at queuesArrivals, service, queue discipline, and capacity shape queueing behaviour. The examples show how assumptions that look similar in a diagram can lead to different performance measures.
  9. 09Stochastic Petri NetsStochastic Petri nets add timing to the representation of states, events, and concurrency. The lesson uses them to describe synchronisation, resource contention, and random transitions, connecting graphical models to quantitative analysis.
  10. 10Separable modelsSeparable queueing networks admit product-form solutions under specific assumptions. This property makes complex networks tractable by analysing their service centres and deriving system-level measures.
  11. 11M/M/1The M/M/1 model combines random arrivals, exponential service, and one server. It makes stability and the relationships among utilisation, queue length, waiting time, and response time concrete.
  12. 12M/M/2 and related modelsThe lesson extends M/M/1 to multiple servers, infinite-server systems, and finite-capacity queues. Comparing variants shows how parallel service, blocking, and limited buffers change congestion.
  13. 13Queues with general service timesThis lesson moves beyond exponential service times to more general distributions. It compares arrival and service assumptions and explains how they affect available solution methods.
  14. 14Introduction to multiclass modelsMulti-class models distinguish jobs with different routes, service demands, or priorities. The lesson explains why class-specific behaviour matters to aggregate system performance.
  15. 15Multiclass open modelsThis lesson applies open-network reasoning to multiple workload classes. Class-specific arrival rates and service demands show why a single aggregate arrival rate can hide important differences.
  16. 16Multiclass closed modelsClosed multi-class models capture fixed populations with distinct service demands or think times. The lesson relates class-level measures to shared service centres and overall response time.
  17. 17Mixed modelsMixed models combine open and closed classes in one system. The lesson explains how external arrivals interact with a fixed user population when both share resources.
  18. 18JMT advanced features for queuesThe final lesson presents advanced JMT features for modelling and studying queues. It brings together formal models and a practical workflow for building a model, interpreting results, and checking assumptions.

ABOUT THESE MATERIALS

Online access
to the slides.

Slides remain in English, the original language of the supplied materials. Introductory commentary is in English; an Italian version is also available. Original presentation files are not included in the website package.

Back to courses →