Grid Computing - Science topic

Can I know more about grid computing ?

Like...

1.Explain risk in your business and how

you can protect yourself against risk?

2.How you can reduce IT overhead?

3.Where are the efficiencies within

technology used in organization?

Goodmorning Monika Dash , ladies and gentlemen, It would NOT be enough to create tiny equilateral triangles to achieve the mesh. Above all, it is necessary to find the optimum surface capable of taking up the spherical shape with the conditions of curvature or geodesic angles. First establish the calculations of the geodesic lengths, then the optimal traingular surfaces according to the size of the circumference of the sphere. We agree that there must be proportionality.

I think this way will make it easier for you to have the constant spring technique in the traingle. The choice of the nature of the triangle, equilateral, isosceles and rectangle must be justified according to their flexibility effiiency.

Maybe you have fixed this problem and you can conveniently share the solution for others who may find themselves in the same situation as you.

Good question! Thanks and good luck.

If you want to increase your definition step by step, you can add a parameter to your array or similar data structure:

-step 1 a[i,j], that's your current description with large cells

-step 2 a[i,j,k,l], that's the higher resolution with smaller cells in each of the previous large cells

within the neighborhood of each (i,j), you introduce further divisions (smaller cells: parameter k horizontally and l vertically)

etc...

In this way you can always go back to less definition.

Does it help you?

Dear Chirag Kharwar ,

Just see this complete and modern hot reference:

Article Cloud IoT as a Crucial Enabler: a Survey and Taxonomy

Dear Adeolu Olabode Afolabi ,

Just see this complete and modern hot reference:

Article Cloud IoT as a Crucial Enabler: a Survey and Taxonomy

Dear Dr. Dheeraj Rane ,

Just see this complete and modern hot reference:

or:

Nidhal El-Omari, “Cloud IoT as a Crucial Enabler: a Survey and Taxonomy”, Modern Applied Science, The Canadian Center of Science and Education, published by Canadian Center of Science and Education, Canada, p-ISSN: 1913-1844, e-ISSN: 1913-1852, DOI:10.5539/mas.v13n8p86, 13(8):86-149, 2019.

I have another paper for the is related suggest an engine over the cloud for Image Processing is:

Nidhal El-Omari, “An Efficient Two-level Dictionary-based Technique for Segmentation and Compression Compound Images”, Modern Applied Science, The Canadian Center of Science and Education, published by Canadian Center of Science and Education, Canada, p-ISSN:1913-1844, e-ISSN:1913-1852, DOI:10.5539/mas.v14n4p52, 14(4):52-89, 2020.

Dear Vishnu,

in which format is the data, which you would like to interpolate, available: NetCDF, ASCII-Text, Excel, ... ?

Don't go with HEC-HMS for filling sinks and other purposes. Do them all in ArcGIS and import that DEM into HEC-HMS. Next create the catchment or sub-basins shapefiles containing the following labels: catchment area, catchment centroid in lay and long. Import this shapefile into HEC-HMS and give its projections. In GIS module of HEC-HMS, follow the required steps and finally create GRID.

The input files are to be given in gridded format only. For that use asc2dssgrid batch file for generation of time series gridded data.

The documentation on the MSH file format states:

"All the node, element and entity tags (their global identification numbers) should be strictly positive. (Tag 0 is reserved for internal use.) Important note about efficiency: tags can be "sparse", i.e., do not have to constitute a continuous list of numbers (the format even allows them to not be ordered)"

See here: http://gmsh.info/doc/texinfo/gmsh.html#MSH-file-format

So better not make that assumption. I would be surprised if nodes and elements in MSH files generated by gmsh weren't in ascending order. However, gmsh can still read and modify MSH files for which this is not the case - this could for example be the case when you're manually editing or creating MSH files.

Avant de calculer ces 2 grilles , il faut préparer et transformer ta grille au domaine de Fourier pour avoir (.FHX et .FHY).

Alors ces deux fichiers (.FHX et .FHY) sont dans le domaine de Fourier , il faut appliquer la transformée de Fourier inverse à ces deux fichiers de sortie.

Many scheduling techniques/algorithms were presented and you should read some survey papers on this.

Read the related works section of following paper:

Also you can check these out:

Grid computing system is a widely distributed resource for a common goal. It is Brother of Cloud Computing and Sister of Supercomputer. We can think the grid is a distributed system connected to a single network. This types of computing work with the large volume of files. Basically, it is a cluster types system. So people call it cluster computing.

Source: https://www.cloudwebhostingtips.com/grid-computing-vs-cloud-computing-supercomputer/

Grid computer tends to be more geographically disperse and heterogeneous by nature. Grid network also has various types. A single grid is like dedicated connection but a common grid perform multiple tasks.

The size of the grid is large. So grid computing is like supercomputing. It consists of many network, computer, and middleware. Grid computer is dedicated to some specific function of the large volume of data. In the grid process, each task divided into a various process. All the process starts execution simultaneously on a different computer. As a result, very few seconds needs to execute and enjoy the flavor of supercomputing.

For more:https://www.cloudwebhostingtips.com/

Grid computing system is a widely distributed resource for a common goal. It is Brother of Cloud Computing and Sister of Supercomputer. We can think the grid is a distributed system connected to a single network. This types of computing work with the large volume of files. Basically, it is a cluster types system. So people call it cluster computing.

Source: https://www.cloudwebhostingtips.com/grid-computing-vs-cloud-computing-supercomputer/

Grid computer tends to be more geographically disperse and heterogeneous by nature. Grid network also has various types. A single grid is like dedicated connection but a common grid perform multiple tasks.

The size of the grid is large. So grid computing is like supercomputing. It consists of many network, computer, and middleware. Grid computer is dedicated to some specific function of the large volume of data. In the grid process, each task divided into a various process. All the process starts execution simultaneously on a different computer. As a result, very few seconds needs to execute and enjoy the flavor of supercomputing.

For more:https://www.cloudwebhostingtips.com/

Grid computing system is a widely distributed resource for a common goal. It is Brother of Cloud Computing and Sister of Supercomputer. We can think the grid is a distributed system connected to a single network. This types of computing work with the large volume of files. Basically, it is a cluster types system. So people call it cluster computing.

Source: https://www.cloudwebhostingtips.com/grid-computing-vs-cloud-computing-supercomputer/

Grid computer tends to be more geographically disperse and heterogeneous by nature. Grid network also has various types. A single grid is like dedicated connection but a common grid perform multiple tasks.

The size of the grid is large. So grid computing is like supercomputing. It consists of many network, computer, and middleware. Grid computer is dedicated to some specific function of the large volume of data. In the grid process, each task divided into a various process. All the process starts execution simultaneously on a different computer. As a result, very few seconds needs to execute and enjoy the flavor of supercomputing.

For more:https://www.cloudwebhostingtips.com/

Grid computing system is a widely distributed resource for a common goal. It is Brother of Cloud Computing and Sister of Supercomputer. We can think the grid is a distributed system connected to a single network. This types of computing work with the large volume of files. Basically, it is a cluster types system. So people call it cluster computing.

Source: https://www.cloudwebhostingtips.com/grid-computing-vs-cloud-computing-supercomputer/

Grid computer tends to be more geographically disperse and heterogeneous by nature. Grid network also has various types. A single grid is like dedicated connection but a common grid perform multiple tasks.

The size of the grid is large. So grid computing is like supercomputing. It consists of many network, computer, and middleware. Grid computer is dedicated to some specific function of the large volume of data. In the grid process, each task divided into a various process. All the process starts execution simultaneously on a different computer. As a result, very few seconds needs to execute and enjoy the flavor of supercomputing.

For more:https://www.cloudwebhostingtips.com/

follower

All Grid and cloud computing platforms have a dispatcher or service broker to manage the distribution of requests. Above diagram explains it well.

Thank you,

Sham

Devang Swami ,

Thank you for your answer ,I appreciate it.

best regard.

Kennedy Chinedu Okafor

Thank you for your answer

can you tell me the right one please ?

best regard.

Hi Sara

Sorry for the delay in getting back to you. The last couple of weeks have been somewhat fraught.

Anyway, my interest lies in all aspects of cloud, whatever the flavour. My initial focus was on key management issues, followed by technical vulnerabilities and forensic issues, which can be problematic. I have tagged IoT into the mix along the way, and will then move on to the the Threat Environment itself. By tackling the problem from different angles, my hope is that a better all round solution can be found. And there remain other issues yet to be properly addressed. This might take some time. Often, companies do not exactly help themselves as much as they could. It will be some time before all the holes can be properly identified, and properly mitigated.

Regards

Bob

maybe i can help you but need more explanation what you want exactly!

Dear Mariusz,

If you can please share your data please. I will acknowledge your contribution in my work too. 

Regards...

Globus Toolkit

As you are simulating practical conditions in Fluent, you should validate your results with the experimental data you have. First you make a grid that produces the same result as that of the experiments did within an acceptable limit (this may be 2-5 % typically). To make sure that your result is grid independent, you change the no. of grids globally not locally. Say you increase/decrease the no. of grids to 10% of your total grids and if the results do not vary then you can say your result is grid independent. You need to perform this at least for five different grids. To optimize your grid you need to find what is the minimum no. of grids that gives you the same results as obtained by the previous simulations within the acceptable error limit.

Maira Abur - Adding to the above papers, you can refer my research work on Intrusion Detection System for Cloud. 

These papers are published in 2015. You can find the latest and current situation of research on IDS via my papers.

Thank you.

Elliptic smoothing can be done using 2D or 3D surfaces, and 3D volumes. For surfaces comprised of structured grids where there is an underlying orthogonal system for the computational coordinates mapping one to one from opposing sides, you can use the methods of Steger & Sorenson or Thomas & Middlecoff. I prefer Steger & Sorenson because you can reconstruct the Poisson equation at each point. This is a boundary value problem where you specify the orthogonal condition, usually with a dirichlet boundary, then solve the PDEs at the boundary for the source terms. From there you can interpolate the source terms onto the interior with TFI or simply decay them from the boundary using bi-linear interpolation with an exponential function to get the source terms in the domain. Then, lagging the solution to linearize the non-linear equations, you can solve the PDE iteratively. You will have to update the right hand side terms at each iteration or use a planar implicit solver. The same holds for 3D, but is far more complex. Check out AIAA 2003-0952 for 3D.

Observe the PV characteristics of your PV. Note the MPPT. In my opinion the voltage is such that Power is zero. "This is not surprising and not a wrong statement". The PV curve is like an inverted parabola so there is a potential where power is zero.( this is a voltage other than zero itself).  More so your inverter input maybe at this potential causing the power to be zero. I would recommend using a dc-dc converter between PV and inverter. The duty cycle ratio should be inverter input voltage:PV voltage. Of your MPPT is lets say 40V, use a boost converter (for example) to give 80v then a inverter followed by a step up transformer. Voltage regulation is also desirable, follow the MATLAB vid on PV-grid integration, by carlos.

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

Hi Tamuno Tobi, just to add to list provided by Jiwan Ninglekhu, the following barriers should be considered:

1. Insufficient Bandwidth Penetration - a successful cloud implementation requires careful attention to bandwidth needs.

2. Deficient Power Supply ( Underpowered) - the quality (consistent availability) and cost of electricity is also a big factor in many developing economies (e.g Africa) since they influence datacentre offerings (i.e. the cost of running generators during power outages).

3. Lack of Internet Coverage - cloud computing’s lifeline is connectivity-meaning all services, benefits, and opportunities that it is envisaged to deliver only exist when we are connected to the network  i.e. no network = no services.

4. Lack of Access Device -  for businesses (especially SMEs) to maximise cloud computing benefits, services should not be limited to those only available through mobile devices but extend to computers. This is true for services offered by IaaS delivery model, where appropriate devices might be PCs not only mobile devices like Tablet PCs, PDAs, or mobile phones. Hence the need to provide these services, as well as the connectivity capacity, is a daunting challenge for developing economies.

5. Inadequate Infrastructure - is one of the key impediments to faster economic growth. Most rural communities in developing economies lack basic infrastructure such as good access roads, telecommunications, electricity and water. Considering the fact that, these infrastructures are the backbone of any development initiatives. This is clearly not helping in the fight to bridge the digital divide. Moreover, the absence of such infrastructures leads to the unavailability of Internet infrastructure and escalating costs of such services and the costs of devices to access the Internet.

Bringing all these factors together, from a business persepctive, the concept of cloud computing adoption may be challenging but not impossible to implement in developing economies.

Yes u can setup a grid computing env in local area network, You can have Eg: 1 head node and 2 Compute nodes. If more compute nodes are added  gromacs or NAMD, Simulation (Execution) will be faster.

you may have a look at this paper :

A grid is a collection of resources. Those resources might be clusters or clusters plus other resources. In general, a grid is a geographically dispersed and more importantly organizationally and administratively diverse collection. I.E. a virtual organization. It is usually about more than just scheduling jobs on compute clusters, although for some grids (such as NSF's TeraGrid) that turns out to be the most important aspect. It can be a mushy idea since it was taken up and used as a marketing term, often by folks you didn't fully grok it. The classic paper as a reference is http://toolkit.globus.org/alliance/publications/papers/anatomy.pdf

I see nothing wrong with using dropbox for this.  But you should be aware that dropbox is not a charity: they'll only let you use services free for quite low levels of activity.  As others have mentioned, there are a variety of other free providers that you might leverage.  Of course, that means you'll find yourself wasting time managing a "farm" of free file stores, rather than getting work done.

I sympathize with the need to get by with low budgets, but I'm not sure you've thought carefully about how cheaply you can do network-storage yourself.  After all, the main value of dropbox is the syncing and platform coverage - if all you want to do is shift files over the internet, dropbox may not be idea.  For instance, there are many, many website-hosting sites that provide unlimited storage and bandwidth.  You may actually find it much more convenient to use https to get and put files - you may even find that some of your information is best stored in a DB (again, cheap hosting often provides preconfigured SQL access.)

Mobile grid computing can be categorized as a peer to peer computing, any peer to peer simulation tools can be used to simulate it. One of the simulation tools example is PeerFactSim.KOM. 

Hi,

I'm not sure I fully understood what you're trying to do but the following code does the steps you've requested. You may however adapt some of them for your intended purpose. The code 1) creates a 512*512 array, 2) selects a bit of it, 3) rescales it to have more array elements, 4) Passes the gaussian filter over it, 5) and finally views all results (with matplotlib).

from __future__ import division

import numpy as np

import scipy.ndimage as im

import matplotlib.pyplot as plt

# 1) CREATING SOME TEST DATA

X = np.linspace(-5, 5,512)

Y = np.linspace(-5, 5,512)

X, Y = np.meshgrid(X, Y)

R = np.sqrt(X**2 + Y**2)

Z = np.sin(R)

# 2) SELECTING A BIT OF MY DATA

x0,x1 = 200,400 # X COORDINATES OF PARCEL

y0,y1 = 400,512 # Y COORDINATES OF PARCEL

Z2 = Z[x0:x1,y0:y1].copy() # CREATING TE PARCEL FROM COORDINATES

# 3) RESCALING THE BIT OF DATA

nx = int(1/((x1-x0)/512)) # APROX. NUMBER OF REPEATABLES FOR X

ny = int(1/((y1-y0)/512)) # APROX. NUMBER OF REPEATABLES FOR Y

Z3 = Z2.repeat(nx,axis=1).repeat(ny,axis=0) # BUILDING A RESCALE FOR Z2

# 4) PASSING A GAUSSIAN FILTER OVER IT.

Z4 = im.filters.gaussian_filter(Z3,3) # SIGMA 3 (PUT WHATEVER YOU WANT)

# VIEW ALL RESULTS

plt.imshow(Z.T,origin='lower',interpolation='nearest')

plt.title('Original')

plt.show()

plt.imshow(Z2.T,origin='lower',interpolation='nearest')

plt.title('Parcel')

plt.show()

plt.imshow(Z3.T,origin='lower',interpolation='nearest')

plt.title('Rescaled parcel')

plt.show()

plt.imshow(Z4.T,origin='lower',interpolation='nearest')

plt.title('Gaussian convoluted rescaled parcel')

plt.show()

Hi,

If network simulation is important, you should definitely consider the SimGrid toolkit (http://simgrid.gforge.inria.fr) that relies on fluid network models whose validity has been thoroughly evaluated (see for example http://hal.inria.fr/hal-00872476) and whose implementaion is often orders of magnitude faster than classical packet-level based simulators (http://hal.inria.fr/hal-01017319).

The Simgrid toolkit ( http://simgrid.gforge.inria.fr ) has been extended to address virtualization and cloud computing challenges. Additional information regarding some examples are available on http://simgrid.gforge.inria.fr/contrib/clouds-sg-doc.html

There has recently been work on using simgrid for cloud computing simulation (see for example http://dl.acm.org/citation.cfm?id=2568524). I was not involved in this work but simgrid seemed flexible enough to model non-trivial performance phenomenon of  VM migrations. 

Knowing little about your phd subject, I cannot tell whether SimGrid is perfectly fitted for you but you should definitely look into it...

data centers house thousands of compute, store, and communicate devices to service users requests and application. While energy efficiency is not a goal, the major business objective of most of cloud providers is to achieve minimum service downtime and efficient response. However, if energy efficiency is a goal, virtualization can be used to limit the number of servers that are powerd on. Workload from under-utilized servers can be migrated and consolidated on fewer number of servers, thus achieving energy efficiency. The overhead to workload migration come in terms of IP changes and storage migration cost etc. 

real-time go with Impala/parquet on HDFS

Analytics fo with Spark/parquet/HDFS

the Grid environment is changing dynamically in time mainly due to load of the system, available network bandwidth and also due to other events such as resource or network failure or recovery.

Speed, elasticity, cost, and support are few factors that you may want to consider. For instance speed lets you know how many MIPS can your task handle on a reference VM instance, say small. From here you can extrapolate to other instance types. Time in seconds can be used instead of MIPS. The size of the task can be used to determine itsmigration time. If tasks are web apps processing throughput can be considered. In this case each task would have a rate at which it consumes and produces messages. Elasticity can be derived fron here. Depending on the incomming messages you may need to add/remove tasks so that messages get consumed immediately. The cost also derives from here but depends not only on how many tasks you have but also on how many VM instances you use to run them. Finnaly certain instances may not support your task. For instance one without a certain library or with the wrong OS.

There is not really enough information to provide you with an accurate answer BUT if the FE model you are using is able to capture the exact solution then mesh refinement is superfluous.  For example, all (most) finite elements can model exactly a constant stress state so if your loading conforms with such a state the solution will be independent of mesh refinement. 

For distributed computing most people will use MPI. If this is the case for you that you will use MPI for parallelization in programming your software you can just install an MPI library on these computers, e.g. MPICH2. In addition to this you probably also want a job scheduler. Have a look for PBS and qsub. There are many compatible implementations of these, most of them providing also free or open source versions. Another alternative for the job scheduling is Condor which also works with Windows. The only problem you will encounter with this kind of setup is that speed-ups will not be very high. The network will be the bottleneck unless you can keep communication in your applications to a minimum.

You should also look at SST, the structural simulation toolkit. The SST/macro tool might be useful in your work. See http://sst.sandia.gov/index.html for more info.

The answer lies in software defined networks and OpenFlow enabled switches. I lot of work has been done in this regard recently.

The software approach you are talking about can be implemented as an application of NOX OS.

Hi Sudhakar,

Have a look at this paper:

Chen, W.-Y.; Song, Y.; Bai, H.; Lin, C.-J. & Chang, E. Y.

Parallel Spectral Clustering in Distributed Systems. IEEE Trans. Pattern Anal. Mach. Intell., 2011, 33, 568-586

MPI is a message passing library interface specification for parallel programming.

MapReduce is a Google parallel computing framework. It is based on user-specified

map and reduce functions

It also says: "In general, MapReduce is suitable for non-iterative algorithms where nodes require little data exchange to proceed (non-iterative and independent); MPI is appropriate for iterative algorithms where nodes require data exchange to

proceed (iterative and dependent).

Best Wishes,

Ashkan

Check out thee links for details about a large grid reliability and stability analysis:

My email address is: sasan.barak@gmail.com

As you say, both grid anc cloud are distributed computation paradigms. Also

cloud computing and grid computing are both scalable. Scalability is acheived through load balancing of application instances running on a multitude of operating. CPU and network bandwidth is allocated and de-allocated on demand.

However, there are several differences, for example:

1) grid efforts often focus on running very large applications in batch mode while the cloud isn't a batch system and has a focus in the web application delivery space as opposed to data centers

2) grid typically requires expert knowledge in setting up and in execution (the end users are at least proficient with distributed computing) while cloud does not require any proficiency at least from the end user (he does not have to care about the mapping of task into resources, etc.)

3) it can be argued that cloud computing can be offered on top of grid computing by offering on-demand resource provisioning (http://www.ibm.com/developerworks/library/wa-cloudgrid/) but this is not always the case

Big Data is quite different since it targets to address specific challenges that arise when data have high volume, variety and velocity (IBM adds also veracity). So if cloud and grid focus on the infrastructure and platform part, big data focuses more on the platform and application domain part. Note that there is an overlap on the platform side and it is also quite possible to have cloud enabled Big Data solutions (for example to provide on-demand access to resources). In this manner they can be complementary (http://readwrite.com/2013/06/07/how-cloud-computing-democratizes-big-data#awesm=~ozQF78RtH19q6K) athough both can also exist on their own.

Inter-operability among various Grid middlewares.

Not very clear. Speed ​​tasks with virtualization. Compare with the performance evaluation of algorithms?

You can get carried away by the experiments, different virtualization different hypervisors, etc. Ask what you want to explore.

Here's one article my students made with me. Maybe it's going to be helpful.

WorkflowSim supports simulation of workflow scheduling.

I would suggest NS3 is more powerful simulator for data centre architecture. But on the other hand it is not that easy to manipulate and person's should have programming skills. on the other hand, CDNSim have GUI support which is a plus point but in my recommendation NS3 is mush better. But at the end choice is your because you know the scenario and your test requirements.

as far as topology itself is concerned, the tree-based ones are still dominant in the existing operational DCs, either single-rooted or multi-rooted (fat-tree). In most cases, they are single-rooted tree with redundant roots, i.e.,border routers/switches running such as VRRP.

Hi Mr.Singh,

First you have to open two .nc files in GrADS (eg: a.nc, b.nc) and let us take one variable pr.

I can give u a sample grads script (sample.gs) for doing composite.

***********************sample.gs******************

'reinit'

*open two files

'sdfopen a.nc'

'sdfopen b.nc'

* pr.1 means variable pr from file 1

* taking variable pr from file 1 (a.nc)

'set time 1jan2002'

'define x1= pr.1'

'set time 1jan2003'

'define x2= pr.1'

'set time 1jan2004'

'define x3= pr.1'

* taking variable pr from file 2 (b.nc)

'set time 1jan2005'

'define x4= pr.2'

'set time 1jan2006'

'define x5= pr.2'

'set time 1jan2007'

'define x6= pr.2'

'define composite= (x1+x2+x3+x4+x5+x6)/6.0'

'set gxout contour'

'd composite'

'quit'

***************************************************

This script may be useful for you.

I am not aware of any grid computing service offered in exchange for money. And I believe there are very good reasons: grid computing developed as a way to share and federate excess resources within a certain community. Therefore, grid security model is based upon the concept of trust. Access to resources is only granted to trusted individuals and services. Trust can not be bought with money.

Cloud computing in my opinion should be used for storage only. While I have reservation about privacy and viability, most IT technicians have the idea that cloud computing can take the place of a true server,,,it does not. Systems such as NAS should be what they are and nothing more. NAS is more cost effective for what it does and nothing more...my thoughts.

Using the term of cloud computing is only marketing hype. Before we had clouds, I used webmail. Now, I am using the exact same thing, even the user interface did not change, but it's called e-mail in the cloud. I used to have my calenders and contacts stored with an on-line service, now they are in a cloud. People used to have FTP servers and the like, now files are stored in the cloud.

Basically, the concepts haven't changed, but now everything is called "to be in the cloud". This means that cloud computing is outsourcing HPC. Though cloud computers are not the same thing as regular clusters since they are usually based on virtual machines.

So, the real question is if companies will outsource HPC or use their own clusters. In some cases it is a trade-off between confidentiality and cost (with cloud computing you only pay what you use). If a local cluster is used heavily cloud computing might not make sense financially. Therefore, I guess that cloud computing is not replacing HPC for everyone.

DICOM Part 3 contains the "DICOM model of the real world" and information object definitions.

Yes. People may share their computer processing power (CPU or GPU) to let other people enjoy parallel and distributed processing. Around ten years back, such a technique had a fancy name called Grid Computing. It was the task Grid Middleware such as Globus, gLite, UNICORE, etc. Examples of fastest virtual super computers are Folding, Bonic, etc.

I suggest to have a look at the following link at wikipedia. http://en.wikipedia.org/wiki/Grid_computing

Similar kinds of things are possible in case of P2P networks or protocols applied to distributed cloud computing.

Kamel Mazouzi : Does GridWay really support workflow execution?

We want to respect two SLA attributes: data transfer time and job execution time.

I.e., if we want to make sure that the data transfer between two Grid sites occur within specified time and job completes its execution on particular Grid site within specified time, how is it possible?

If we want to estimate for some real system, not on a simulator, is it possible to estimate it?

“A multiobjective mesh optimization framework for mesh quality

improvement and mesh untangling”, Jibum Kim, Thap Panitanarak, and Suzanne M. Shontz, Int. J. Numer. Meth. Engng 2013; 94:20–42. Maybe this paper can help you.

What do you mean? The behavior of any batch system is knowable exactly as much as you know the distribution and properties of running and queued jobs. Do you know the actual runlength of the jobs? Do the jobs have differing resource requirements (cores-per-process, memory-per-process, processes-per-node, etc)?

There are a number of hot issues in Cloud Computing, as mostly listed down in the following paper. In my view, the research issues that need immediate attention include data confidentiality and auditing.

Hi,

According to the RSL specification (http://www.globus.org/toolkit/docs/5.0/5.0.0/execution/gram5/pi/#gram5-rsl-attribute-summary) I would say it's not possible. You can easily do/check a reservation by executing remotely qrsub. The real problem is how to pass the AR number with GRAM. The GRAM-SGE perl module is a helper around (qsub, qdel, qconf).

platform as a service from open source community and redhat: openshift

this p-a-a-s software is included in linux distribution fedora 19.

OpenStack is one of the best available option for now....as it open source and freeware we can do all the type of R&D upon it. Its implementation is little bit hard but we can try to understand it .. here is the refrence link ...http://www.openstack.org/

Globus toolkit is widely used Grid middleware for implementing Grid computing concepts. You should visit http://www.globus.org/

I suggest some articles:

If by "print" you mean publish, an alternative can be grouping those instances in OWL files.

Before you ask what sections should be contained in a survey paper, you should first understand what is a survey paper. What it is *not* is simply a core dump of a bunch of papers in a common area.

Think of a survey as a research paper whose data and results are taken from other papers. This means that you should have a point to make or some new conclusion to draw. And you'll do so by collecting data from a broad collection of previous works.

As a previous responder has said, you should have a thorough and deep knowledge of the field that you are surveying. This knowledge should be sufficient to be completely aware of the the main themes, directions, controversies, and results in that field.

The point you will make will determine the organization of survey paper. The structure of the main sections of the paper will reflect the structure of field. Some possible example structures (which of course depend completely on the topic) might be:

1. Increasingly complexity or scale: There may be a spectrum of solutions and you might organize them by complexity or scale.

2. Static vs. dynamic: Many field organize by static techniques, dynamic techniques, and even hybrid.

3. Partitioning the design space: Lots of systems are made up of components, so maybe for an compiler paper, you could divide by the classic scanner, parser, symbol table, code generator and optimizer.

4. Major techniques in a field: For example, in fault tolerance, you see fail stop vs. fail forward, or logging vs. hot-backup. In concurrency control, there is a natural divide between optimistic vs. pessimistic techniques.

5. Historical: sometimes the course of development of a field has a clear linear nature and is intrinsically interesting in itself. This is an over-used techniques in many cases where it really doesn't add understanding.

There are lots of possibilities for a given topic and it is this organization that is the hardest part of writing a survey paper. (I'm sure that many of you can give good examples of organizations that have worked well for you.)

You'll have written a successful survey paper if you can communicate not just the list of results, but more important, your understanding of the structure of the field.

This is a high bar to set. And it is also why I never ask students in my graduate classes to write such papers; they just don't have the experience and perspective to write a good survey.