Projects

Energy-efficient computer system for high performance computing





NCN grant (OPUS-9, No 2015/17/B/ST6/01885)
Granting period: 2016 - 2019

Project contractors:

• Ewa Niewiadomska-Szynkiewicz (coordinator)
• Michał Karpowicz
• Piotr Arabas
• Michał Getka
• Antonina Krajewska

This research project addresses the vital problem of energy efficient high performance distributed and parallel computing. The main scientific challenge is to acquire new knowledge on the stochastic dynamics of data processing in HPC systems and to propose adaptive resource management algorithms efficiently exploiting new power control capabilities of hardware elements.
The research objectives are targeted on providing contributions to development of future generation computing and operating systems. Specifically, the objectives include characterization of energy-efficient service rate control laws, identification of computing dynamics at the hardware and operating system level, and development of fast task scheduling algorithms (NP-hard problem). The open question being addressed is how to apply machine learning techniques to exploit dynamic voltage and frequency scaling (DVFS) and smart stand-by mechanisms, as well as high resolution hardware monitoring sensors, to optimize HPC operations both on cluster-node and cluster-wide level.
The project will address the following hypotheses:

• H1. Workload of servers, processors and cores can be viewed as a self-similar stochastic process and predictive power of the best linear approximation of server workload is sufficient to provide quality of service guarantees for energy-aware performance control.
• H2. Optimal energy-aware CPU frequency control policy is bounded from below by the short-term energy-aware application-specific demand tracking (best-response) rule.
• H3. Optimal power budgeting and job assignment provides stability margins for the data center operations meeting service bounds at minimal energy consumption cost.

The proposed approach will result in development of solution that will prove useful from the viewpoint of HPC service development. To verify this claim obtained outcomes will be subject to formal and experimental verification in laboratory.

• 
  National Cybersecurity Platform
  

  
  RESEARCH PROJECT - NCBiR
  

  Granting period: 2017 - 2020
  Coordinator: Research and Academic Computer Network - National Research Institute (NASK-PIB)

  Project contractors:

  • Ewa Niewiadomska-Szynkiewicz
  • Michał Karpowicz
  • Piotr Arabas
  • Adam Kozakiewicz

  The goal of the Project is to develop a comprehensive, integrated system for continuous monitoring,detection, and warning of threats identified in a near real-time in the State’s cyberspace. A prototype of a National Cybersecurity Platform (NCP) comprised of an Operational Centre (OC) and components that integrate participants of the NCP with the OC will be the main outcome of the Project. The NCP prototype, proven in operational environment, will provide nationally coordinated actions to prevent, detect and mitigate the impact of incidents that violate the security of ICT systems vital to the functioning of the State, and will create opportunities for sharing cyber security awareness within the European Union. System and technical solutions, developed within the NCP Project, will enable collecting the current information on the cyberspace status, assessment of the negative impact of identified threats on functioning of the State and undertaking coordinated actions at the national level. The novel methods and tools will be developed with focus on: cyber events correlation, situation awareness evaluation, static and dynamic risks assessment, multidimension visualization with multimodal user interface, vulnerability and threats detection in IoT, TCP/IP, mobile wireless and industrial networks. The advanced access control and dissemination mechanisms for secure information sharing within the NCP will be provided. The part of the NPC will be an expert system supporting the choice of methodologies and metrics to establish ICT security requirements and reporting significant incidents that violate the information security of the infrastructures encompassed by the NIS Directive and future low regulations in Poland. The Project final result will rise the security level of the Polish cyberspace, increase the national competence in cybersecurity and, ultimately contribute to the growth of the Polish economy.
  

International grants
• 
  High-Performance Modelling and Simulation for Big Data Applications
  Projekt cHiPSet
  
  


PAST RESEARCH GRANTS


International grants
• 
  ECOnet - low Energy COnsumption NETworks (ECONET)
  
  www.econet-project.eu
  
  
  
  (FP7 ICT-2009.1.1)
  Granting period: 2010 - 2013
  The consortium consists of 16 partners (universities, research institutions and commercial firms)
  Project coordinator: Consorzio Nazionale Interuniversitario per le Telecomunicazioni (Italy)

  ECONET aims at studying innovative techniques and architectural solutions to support energy efficiency in next generation networks. The ECONET project will focus its research and development efforts in three main research axes. In the first axis, novel network-specific HW/FW technologies will be developed to optimize the power management features. The second axis will be devoted to develop local and distributed frameworks for dynamic optimization of the trade-off between energy consumption and network performance. The last axis will focus on the design of a Green Abstraction Layer for interfacing the novel low-level green capabilities with OAM frameworks in a common and standard way. The ECONET project will deliver novel energy-aware device prototypes on which large-scale demonstration tests will be conducted. The project will aim at maximizing its impact on industrial and network operator communities as well as on standardization bodies.

• Quality of Service and Pricing Differentiation for IP Services (QOSIPS)
  

  (FP5 IST-1999)
  Granting period: 1999 - 2002
  The consortium consisted of 5 partners (universities, research institutions and commercial firms)
  Project coordinator: Knowledge Support Systems Ltd. (England)

  The objective of the QOSIPS project was to develop innovative technologies for supporting Quality of Service (QoS) management, service differentiation and price setting of Internet Protocol Network Service Providers. The consortium designed, implemented, tested and evaluated an integrated system (software, hardware and methodology) that performs the following functions:

  1) Provide exhaustive, non-intrusive and accurate measurement of real-time QoS of
  user’s traffic packets without injection of test packets in the network.

  2) Support customer’s profiling (through classification of customer “real” traffic and
  use of QoS oriented services) as an innovative way of generating service
  differentiation.

  3) Support the pricing of QoS oriented services such as QoS oriented Service Level
  Agreement, per application or/and per destination prioritisation, pay-per-use, pay-
  per-class of service, or global prices packaged per customer type, etc.

  The results achieved were fully consistent with the planned goals. The QOSIPS system as a whole promises to become a powerful commercial tool, and first evidence of the business case has been confirmed by the user. The project has completed very successfully. The original project objectives were totally fulfilled .

National grants
• 
  Methods and tools for ad hoc Network design and control
  
  
  
  

  NCN grant
  Granting period: 2011 - 2014
  Complex Systems Group

  The project focuses its research and development efforts in two types of ad hoc networks: WSN - Wireless Sensor Networks and MANET – Mobile Ad hoc Networks. The objective of the grant is to develop new methods, algorithms, software tools for simulation and laboratory equipped with sensors. We focus our research on localization systems, energy efficient communication, and network nodes mobility modeling.

Parallel and distributed global optimization algorithms for large scale systems.

MNiSW grant
Granting period: 2008 - 2010

The research was concerned with high performance computing (HPC). The general objective of the project was to develop, implement and test novel optimization methods. The designed and implemented solvers was applied to solve real-life problems such as control of complex physical systems. Due to the complexity of the considered problems the attention was focused on parallel and distributed computation and issues associated with reduction of computer memory usage. Particularly the research was addressed to: fast and minimal storage linear and nonlinear continuous optimization solvers, hierarchical methods applying various approaches to problem decomposition, deterministic and stochastic global optimization and algorithms applying interval arithmetic tools. The project addressed theoretical investigations, computer implementation of developed numerical algorithms and simulation experiments.

MNiSW grant
Granting period: 2008 – 2010
The consortium consisted of over a dozen partners (local research institutions)
Project coordinator: National Institute of Telecommunications

Carried out work related to issues of development of ICT networks and services such as next-generation technologies, applications and market issues. The aim of the project was to prepare the tools and procedures for implementation of solutions needed for the development of state and local government and having a chance to be competitive in the market. In particular, attention has focused on identifying and constructive analysis of the directions of the evolution of network architectures and protocols underlying the operation of NGN and their design methodology.

The general objective of the grant was to develop and implement the Crisis Management System (CMS) dedicated for urban agglomeration of Warsaw. The results were: a set of threat models (e.g. predictive) and algorithms covering threats defined in the catalogue of urban threat, a demonstrable distributed software components of CMS for threat analyses supporting. A real urban threat was described by: a type of threat, a source of threat, critical infrastructures, possible losses, methods of counteractions, etc. The following type of threads were considered: military, chemical, biological, radiological, fire, flood, network infrastructures (service), terrorist, environmental catastrophes. The focus was on the synergy effect of complex threats. Due to the complexity of the system the distributed software environment was proposed as a simulation framework. The goal of ICCE team was to realize the task of the project: Prediction and simulation of floods of the Vistula river and crisis management in Warsaw during flood.

Own research project KBN
Granting period: 2001 – 2003

Own research project KBN
Granting period: 1998 – 2000

Own research project KBN
Granting period: 1996 – 1998

Own research project KBN
Granting period: 1995 – 1997

The project ordered by the Polish Oil Concern in Plock
Granting period: 1999

Granting period: 1996 – 1999

Granting period: 1999 – 2000.