Distributed Systems

Distributed Systems

  • Rajesh P Barnwal added an answer:
    What are the best books for teaching distributed systems to computer science students?

    I would like to find some updated, teaching-oriented books in the subject...Thanks in advance...

    Rajesh P Barnwal

    In my opinion, "Introduction to Distributed Algorithms" by Gerard Tel from Cambridge University Press is one of the best teaching-oriented book for distributed system.

  • Qi Su asked a question:
    Can anyone suggest some recent papers regarding deployment of content servers for CDN?

    Can anyone suggest some recent papers regarding deployment of content servers for CDN?

  • Sunkara veerababu Viresh asked a question:
    What is right way of Node selection in Radial distribution system?

    What is right way of Node selection in Radial distribution system?

  • Mohammad Hossein Homaei added an answer:
    How can I validate the architecture for internet of things (IoT)?

    I am exploring some research ideas about architecture for IoT. At present I am come up to a naive idea for IoT architecture, but I am not clear about how to validate the naive architecture for IoT.

    Can anyone guide me in this regards, how to validate the architecture for IoT? what tool to be use? or any other guide line?

    Thanks in advance

    Mohammad Hossein Homaei

    This landscape of architecture in IOT from device to application.

  • Mayank Maheshwari added an answer:
    Does anyone know about power loss calculation in distribution network?

    How come the real power loss for the radial distribution system can be calculated from this formula?


    Can anyone tell me the derivation?

    Mayank Maheshwari

    See the below paper it may help in your question.


  • Ricardo Barcelona added an answer:
    How integration of PV system in a conventional distribution system is possible?

    PV system and distribution system are easily modeled in MATLAB simulation but how can we integrate them and what are the useful control strategies to make system balance and how it affects local distribution grid.

    Ricardo Barcelona

    You may learn from the example of Spain that I highlighted in my paper "Wind Power Deployment - Why Spain succeeded?" published by the Renewable Energy Law and Policy Review in September 2012. For PV Solar, a similar consolidation or integration into a renewable energy sub-system could improve the management of the availability (or intermittency) of PV Solar and Wind power. Managed as a portfolio, the subsystem is better able to complement the individual volatilities in volumes that are not correlated that smoothens the peaks and troughs relative to what individual suppliers experience. This proved successful with wind power in Spain where the record supply from intermittent supplies (mostly wind) exceeded 65% of the total at a given period without adversely affecting network security in Spain. These levels of supplies has been reached more frequently over the past four years. For PV Solar, you need to consider some form of "smart" metering where inflows from the grid are measured accurately (this is what traditional grid does), while also able to measure the outflows from the household of power fed to the grid when the household is not using the power generated from the solar panels. These innovations are more readily adapted in developed or efficiently managed power networks. In markets where network infrastructures are weak or under-developed, strengthening the basic infrastructures and grid management is needed before they can apply a similar system that is proving successful in Spain.

  • Cong Liu added an answer:
    Are there any references regarding calculating concurrent transitions from an arbitrary Petri net?

    I mean an approach to find all concurrent transitions directly from a Petri net structure. 

    Cong Liu

    Thanks Djamel, I will read them later.

  • Mariusz Janiak added an answer:
    What are open-source frameworks implementing publish–subscribe approach over IP/UDP, dedicated for small embedded systems?
    There is a commercial available implementation of RTPS protocol, a RTI Connext Micro, which is well scalable for small embedded systems with microcontroler. There is also open-source project ORTE (http://orte.sourceforge.net) that implement a RTPS protocol, but I haven't found its port for microcontroler based systems.
    Mariusz Janiak

    We are basing on the standard Ethernet hardware that is available in PCs and microcontrollers. On the MAC  level, the RTnet implements software TDMA discipline -- Time Division Multiple Access so the arbitration problem has been already solved. What we looking for now is a midlleware protocol that implements the pub/sub paradigm, and is able to operate over the UDP/IP. Our PCs has 1GB NICs but microcontrollers has only 100MB.  Most of our system typically operate with 1ms period so latency should be no more then the tens of us.

  • Akhilesh Mathur added an answer:
    What will be the basic criterion for designing a "Fault current limiter" in distribution system?

    If I want to place a fault current limiter (FCL) in the distribution system, then how can I choose a size of FCL or design it and how can I found exact location where I can locate it in the system...

    Akhilesh Mathur

    Thank you Dear Nikolay Raychev for this useful link....

  • Dražen Tubić added an answer:
    How can I implement DG in matpower to calculate power flow, MATLAB?

    I need to implement randomly generated DG (size and location are random) into distribution system. For example, if i want to implement DG in bus 4 and with the size of 1 MW, what do i need to do to succefsully perform power flow after that? I tried to do that with matpower case14 system, in which generators are by default on buses 1,2,3,6 and 8. I created new row in generator information (mpc.gen) and provided all information needed (real power output, bus number, voltage magnitude) and when i do power flow i get all kinds of errors. I even tried to set bus 4 as PV, and that didn't help. What am i doing wrong and how can i do that? I am using matpower 4.1

    Dražen Tubić

    Thank you J.Vara, i decided to go with Muhammad's solution, subtracting the DG value from loads (their real power demand), which is very simple and gives the same results.

  • Naoufel Boulila added an answer:
    Is there any survey or literature review paper which describes the transformation of embedded systems to Cyber-Physical Systems?

    I am currently studying the relationship between the cyber-physical systems and its predecessor, embedded systems. I am look for survey papers, literature reviews and classical references.

    I am giving special concentration to the security aspects.

    Naoufel Boulila

    i think this book is the right one:

    Embedded System Design: Embedded Systems Foundations of Cyber-Physical

    there is a publication or slides about it here on Researchgate: http://www.researchgate.net/publication/234125635

    + 1 more attachment

  • Hemakumar Reddy Galiveeti added an answer:
    Can someone provide me with the technical data sheet of the RBTS bus system?

    Can someone furnish me with the technical data sheet of the RBTS bus 1, bus 2, bus 3, bus 4, bus 5 and bus 6 transformers i.e Transformers; impedance, ratio etc. I need the data to estimate distribution system technical losses.

    Hemakumar Reddy Galiveeti

    You can get transformer impedance values from manufactures name plate details.

    Ratio depends on the primary and secondary voltage and transformer tap changing ratio depends on the system operation.

  • M. Hamiz added an answer:
    Which DBMS support distributed databases more as compared to competitors?
    Distributed Databases are one of the major requirement of the industry now a days. So would you share your knowledge about those DBMS which support the maximum functionality and requirement of Distributed Database.
    M. Hamiz

    True. Put the fragmentation or replication to the highest demand site

  • Sureshkumar Sudabattula added an answer:
    How the DG penetration plays an important role in distribution system?

    i am searching so many papers in this area some the researchers simply taken an assumption of 100% penetration and adding an distributed generation sources and minimizes the power losses to a very small value but how it influences the existing  distribution network they didn't considering. finally my doubt is what is the maximum percentage of DGs should be connected to already existing distribution network?

    Sureshkumar Sudabattula

    i am expecting that maximum of 50% is taken from distributed energy sources that is only capable for already existing distribution network and also consider the cost these sources taken into account.

  • Holm Xie added an answer:
    Any tool suggestion for visualizing a geographical data distribution of distributed system?

    I would like to visualize the geographical data distribution of a distributed storage system. Can someone suggest me a tool which does this?  

    Holm Xie

    You can try R or Python. They both have plenty package to interworking with map.

    Please check the textbook: Data Driven Security, Analysis, Visualization and Dashboards, Wiley 2014, Jay Jacobs and Bob Rudis

  • Morteza Seddighighachkanloo added an answer:
    How can I performance a load scaling in PSSe?

    How can I performance a load scaling in PSSe?

    Morteza Seddighighachkanloo




  • Closed account added an answer:
    Any advice on the number of data for big data?

    hi guys,  i am working on dataset with 1M user and 943,347 item , but i want to consider less than, for example 10,000. my platform is weka.  my reason is that ram of laptop can not process on 1M. i need a strong reason that why i consider less of data?

    My work is not on big data. i want to know this number (1M) is big data, or need parallel systems or distributed systems? thanks


    Big Data not necessarily mean huge volume of data, if you have varied data <variety> , and rate at which you get those data is also a criteria<Velocity>.

    Well if you want to process 1 million data, you can still do that using virtual cluster offered by Hadoop for which your notebook would be suffice.

  • John Frederick Chionglo added an answer:
    How can I create a Petri net simulator?

    Hello everyone,

    I would like to make a simulation of a particular Petri net.

    Is there someone that can help me with some ideas or principles on this?

    Thank you in advance!

    John Frederick Chionglo

    Some Principles and Ideas

    Consider “making a simulation of a petri net” to be the same as “writing a computer program using petri nets”. To use petri nets for writing a computer program is to organize a computer program using Petri Net elements and annotations. To create this organization, first establish the relations of computations in terms of places, transitions, inputs and outputs. Second create annotations of places, transitions, inputs and outputs for the computations.

    An Example

    Consider making a simulation of a countdown timer. The countdown timer begins from a specified value (say, 10). Using some interval, the timer value decreases by one. When the timer value reaches zero, the countdown ends.

    This computation may be modeled in terms of a place, an input and a transition. The place has a mark annotation, a value that represents the timer value – the place is the timer. The input has a “fire” annotation, a computation that deducts one from the input place mark annotation. The input also has a status annotation, a ‘true’ value means the computation may proceed and a ‘false’ value means that the computation should not proceed. The input has an “is enabled” annotation, a computation that determines the value of the status annotation – in this case, if the mark annotation of the input place is greater than zero, the status annotation of the input is ‘true’ and ‘false’ otherwise. The transition has “fire” annotation, a computation that delegates the “fire” computation by “firing” its input. The transition has an annotation status, a true value means that the value of its input status annotation is ‘true’ and ‘false’ otherwise. Consider too visual annotations of the net elements: a circle for the place, a square for the transition and an arrow from the circle to the square for the transition, and dots for the mark annotations of places – empty for 0 mark, one black dot for 1 mark, two black dots for 2 marks, etc. [The types of annotations considered so far are those found on Place/Transition nets, a kind of Petri Net.]

    To make the “simulation” interactive, consider an event annotation for the transition, a computation that delegates a user event (such as a mouse up event) to “fire” the transition. Furthermore, consider a visual annotation for the status of the transition – green if the status annotation is true and empty otherwise.

    Based on this example, I created an interactive “simulation” of this system in PDF. Graphics were created using PowerPoint and exported as PDF. Computation logic was implemented as JavaScript programs. And interactions were integrated with the computation logic using the Acrobat/JavaScript API. [See link attached to this reply].

    A Reference

    For petri nets with many net elements and annotations, several other issues (such as naming conventions, high-level graphics conventions, and software engineering workflow processes) must be addressed. Here is a research paper related to this topic – Net Elements and Annotations: Computations and Interactions in PDF. [See link attached to this reply].

    + 1 more attachment

  • Arundhati Sahoo added an answer:
    How can we decide the rating of DVR placed in a distribution system for voltage sag mitigation?

     What may be approximate cost of DVR?

    Arundhati Sahoo

    the simulation of a DVR is done using MATLAB/SIMULINK software. Thus it became
    easier to construct the large distribution network and analyse the various result for two different types of faults.
    The controlling of DVR is done with the help of PI controller. The simulation results clearly showed the
    performance of the DVR in mitigating the voltage sag due to different fault conditions in distribution systems.
    DVR is one of the fast and effective custom power devices. DVR has shown the efficiency and effectiveness on
    voltage sag compensation hence it makes DVR to be an interesting power quality improvement Device. This has
    been proved through simulation and hardware implementation. From the analysis it is found that in case of a
    three phase fault almost 91% of compensation is done and in line to line fault voltage compensation took place
    for almost 44%.

  • Eduard Babulak added an answer:
    Are there Petri Net diagrams worth adding "new lives"?

    I would like to add “new lives” to Petri Net diagrams that were published between 1962 – 2014. The Petri Nets with “new lives” would combine the published Petri Net diagrams from the past with JavaScript codes and supporting graphics to create interactive and dynamic Petri Nets in PDF. In other words, the revival creates token game versions of the Petri Net diagrams in PDF.

    I am limiting the number of Petri Nets to a maximum of two per year.

    Question 1: Which Petri Net diagrams should I include? Why should I include them?

    Once I have finalized the list of Petri Net diagrams to revive, I am hoping to finish the work as soon as possible. So before I begin, I will be looking for volunteers who are interested in adding “new lives” to the Petri Net diagrams. I will  create at least two token game versions. Thus I will be looking for a maximum of 102 volunteers – one person per token game version. If there are less than 102, I will create the difference.

    Question 2: Would you be interested in helping out? If so, please give me a shout.

    - john

    Eduard Babulak

    Hi John,

    Subject to your case scenario and system dynamics the embedded Petri Net System may  be relevant.

    Best regards 

  • Nana Cne added an answer:
    Does anyone have data on the workload on Grid/Cloud/P2P or client/server application?

    Please I need a help in emergency,

    I would like to ask if anyone has a benchmark, or workload data on Grid/Cloud or Peer to peer systems, or any client/server applications, to provide me with these data and to allow me to use them for academic purposes, to make some experiments on these data. 

    Any help will be acknowledged, or even best, if some one is interesting by sharing the data in a collaboration form...

    I will be very very grateful..


    Nana Cne

    Thank you Joan Navarro, these are interesting benchmark generator, hoiwever I need a real benchmar or data set.

  • Shivananda Pukhrem added an answer:
    How can I use distributed model predictive control for Voltage Control?


    I am looking to use MPC for voltage control in distribution systems with the existence of DGs and storage units. I am trying to use OpenDss to run power flow while using Matlab to build the control model. My question is, How can I transfer information between the two MPCs and send the control signals to openDss to change control parameters such as LTC, DG power factor,...so on?. Since the openDss solves either snapshot or hourly, Is there any better software to use for simplicity?


    Shivananda Pukhrem


    I have done similar work to interface OpenDSS and Matlab for DG impact on grid. if you are familiar with the discussion forum in OpenDSS, you can raise your doubts. But again from my experience, any algorithms can be implemented through Matlab in conjunction with OpenDSS to solve your network.

  • Danh-Tai Hoang added an answer:
    How can we measure the order parameter (degree of synchronization) for a system with bimodal distribution, according Kuramoto model?

    I am considering a synchronization system of oscillators with Kuramoto model. Under a certain condition (parameter), the phase distribution of the system can be bimodal with a different phase of \pi, and also can be multimodal under other condition. Order parameter as show in attachement is not consistent because R should be 0. Do you know how to measure the degree of synchronization for these cases? Thank you very much in advance.

    Danh-Tai Hoang

    Dear Prof. Ning Cai. Thanks a lot for your suggestion. You are right! The quadratic index as shown in the following equation is consistent with this system.

  • Eric Verhulst added an answer:
    Qualification of a Time and Space Partitioning (TSP) kernel: how formal methods could help ?

    I have an issue concerning the verification and validation of a TSP kernel... I would like to verify that for any configuration of the kernel, the kernel will satisfy the time and space segregation of the partitions. To be precise, I don't want to prove this property for all configurations but only configuration which respects the usage domain fixed by the kernel provider (bounds of the parameters or other structural constraints). The way to verify this is the following is to use unitary tests… To reduce the combination of tests, we have to make assumptions on the independence of (some, but more than less…) variables/configuration parameters. Using these assumptions we can drastically reduce the number of tests to be executed. Remain one thing: we have to prove or provide evidences that our assumptions are correct. To prove it, classical way is source code/ design inspection.

    My question is the following: how formal methods could help me at this stage? Can we prove independence between variables in a source code using formal methods and if yes, which method/technique?

    At first glance I see two ways to be explored:

    -          Either using formal static analysis to help the evaluator in the inspection of the source code and to prove independency of some variables… Have you any references of idea for helping doing that?

    -          Or trying to prove some invariants on an abstraction of the source code.  Does the independence of variables could be expressed as an invariant? How can I obtain a “good” abstraction of my source code to prove my invariant?  I hope that I can use source code annotation (something like ACSL annotation and Frama-C toolset) but I have not enough experience to see if it a reasonable or even feasible idea…. I can also hope that a magic static analysis tool could automatically generate an abstraction of the source code that I could use to prove my invariant using a theorem prover. But as for the previous one I have not enough experience to say if it is reasonable…

    Any idea and reference papers on the subject are welcome…

    Eric Verhulst

    Notwithstanding the (correct) remarks made above on the definitions, I also acknowledge that this is a forum and not a scientific paper. Given the Airbus context, it is clear that Alexandre is talking about ARINC 653. And TSP as defined in ARINC 653 is ARINC 653 specific. But generally speaking, one can see the general of a OS that strictly enforces that applications will only run in their allocated time slot and will never access memory from another application. Given that we speak of timeslots, it is clear that we no longer assume a strict real-time behaviour (unless the time constraints are >> than the sum of all timeslots). 

    However, the question is very interesting. From my point of view, for three reasons:

    1. A formal analysis (formal verification comes later), would be beneficial in defining this TSP much better. The reality is that ARINC 653 and hypervisors are ad-hoc, heuristic solutions that came into being because there was a need for something like TSP. Or rather the Requirement was non-interference/isolation/preventing error propagation/etc. The question becomes: are ARINC 653 and hypervisors goos solutions for this Requirement?

    2. We took a different approach by doing away with timeslicing and using RMS (priority based preemptive scheduling) and using time related scheduling as constraints on top. See VirtuosoNext on altreonic.com

    3. We have now the same question as Alexandre. While the kernel was formally verified, this fine-grain TSP was not and we also don't really know how to do it. Part of the issue is not so much the modified scheduler (with some code generation) tools, but any TSP implementation relies on functionalities offered by the hardware (MPU or MMU). Also, time as such is not a parameter in the scheduler implementation. By doing away with the timeslicing, time and "space" are orthogonal issues.

    Hence, I think formal methods (verification like with TLA+ or proving like with theorem provers) can be applied to the logic of the source code. Pre-condition: this source should assume nothing about time. Formal methods are rather weak dealing with time. If needed, better treat it as a globally shared variable. 

    So what rest is testing to validate the implementation. I would focus on the "border" cases (that might be expressed as invariants in the formal models). How does the system react when a fault is injected? What happens if the CPU load is a bit too high from time to time? Then Alexandre asked about if the tests should cover any possible configuration? I think not because that is non-exhaustive. But is the implementation (and the associated execution time) sensitive to e.g. the number of Tasks/Processes? If so, this can be considered as a design error (one should use datastructures whose access time do not depend on how many items it holds).

    Conclusion: interesting question. More formalisation in this domain would be very helpful

  • Iwona Grobelna added an answer:
    What can i do to rectify the problem of hanging in TimeNet tool for Petri Net,anytime i want to simulate my model it always hang and have to close it?

    Anytime i want to perform token game or transient simulation or stationary simulation and generate graphs in result monitor it always hang, plss i need solution?

    Iwona Grobelna

    For tokengame, analysis or coverability graph generation of Petri Nets I use the WoPeD tool (http://woped.dhbw-karlsruhe.de/woped/). I works quite stable and so far I have no problems with it.

  • Akinwale Akinwunmi added an answer:
    how can a programming language be selected for developing distributed systems?

    Programming languages and distributed systems have long influenced each other. Naturally, every programming language has its strengths and weaknesses. Consequently, it might be difficult to decide precisely which language should be chosen for a software project. However, the selection of the right programming language can be crucial to the success of a project or a software system.

    Akinwale Akinwunmi

    There are programming languages with capabilities for distributed computing such as java and python. The functionality required for such project will determine which way to go after proper scrutiny of the features of the supporting programming languages.

  • Suman Bhullar added an answer:
    In a load flow of radial distribution system can we generate complex power so that current will become minimum and hence maximum V1?

    I have to develop a load flow model of balanced radial distribution system. I want to apply optimization techniques, so that current becomes minimum and voltage at the receiving will become maximum.

    V2= V1- Ijj Zjj

    where V1= voltage at the sending end node and V2= voltage at the receiving end node;

    Ijj= branch current and Zjj= branch impedance;

    branch current further depends upon complex power and voltage which is assumed to be constant.

    By optimization technique can we generate complex power so that current will become minimum and hence maximum V1?

    What will be the objective function?

    F(x)= Min Ijj?

    Am i right?

    or F(x)= Max

    Suman Bhullar

     Thank u to all of u for sparing time for my query. it is very good platform for researchers to discuss their problems .

  • Petros Aristidou added an answer:
    Can anyone suggest an algorithm for the application of MPC in distribution system?


    Can Anyone suggest an easy implemented  algorithm for the application of using Distributed Model Predictive Control for voltage control in Distribution system. A distribution system with Dgs, capacitor bank, storage units, AVR. And is there any way to transfer info between two control models each one controls a specific zone in the system.


    Petros Aristidou

    For more power system oriented applications, check out the following papers. There are many things in their references.

    • [Show abstract] [Hide abstract]
      ABSTRACT: This paper presents a centralized control scheme to regulate distribution network voltages in the presence of high penetration of distributed generation. The approach is inspired of Model Predictive Control in order to compensate for modeling inaccuracies and measurement noise. The control actions, calculated from a multi-step optimization, are updated and corrected by real-time measurements. The proposed controller uses a linear model to predict the behavior of the system and the optimization is solved using quadratic programming. The proposed corrective control has been tested in a 11-kV distribution network including 75 nodes and hosting 22 distributed generating units.
      IEEE Transactions on Smart Grid 12/2013; 4(4):2152-2161. DOI:10.1109/TSG.2013.2246199

    + 1 more attachment

  • Allan E. S. Freitas added an answer:
    What are recent real time issues in distributed system security?

    I want to know the issues around distributed system security and how can be remoduled to give proposed possible things to overcome?

    Allan E. S. Freitas

    You have to keep control on available resources, in order to keep the required constraints for the real-time distributed system.

  • Mostafa Eidiani added an answer:
    Can we draw a simulink model of a radial distribution system?

    I haven't seen any paper containing simulink model of radial distribution system  except some software .e.g. ETAP.

    Mostafa Eidiani

    You can see also this link


About Distributed Systems

Distributed computing is a field of computer science that studies distributed systems. A distributed system consists of multiple autonomous computers that communicate through a computer network. The computers interact with each other in order to achieve a common goal. A computer program that runs in a distributed system is called a distributed program, and distributed programming is the process of writing such programs.

Topic followers (14,750) See all