Friday, July 22, 2005

Average Hop Count of Platform

Now the average hop count from user to platform can be shown in simulator.xls.

Algorithm:

each the average hop count is calculated from each platfrom.

* Platform find the shortest hop count to user location. (in case there are many users).

Average platform hop count = Sumation of hop count of each platform / number of platform



Task:
Change agent migration policy to use healthy level

Thursday, July 21, 2005

July 21st

Simulator Update


Now, we can simulate

8x8 grid network static:

I believe the simulator can show 10x10 grid network.






* We can see in the picture that there are two users in node 0 and node 5.

Agents are divided to two groups to follow each user.

Agent Migration Policy:

value1 = getNodeRequestRateRatio(nodeId);
value2 = getNodePopulationDensityRatio(nodeId);
value3 = getWandering();

if ((maxScore > migrationThreshold ) && (this.energyLevel_>2000) ) return maxScoreNodeId; else return -1;


Task:

1] Change migrati0n policy to use healthy level instead of agent population
2] Number hops count between platform and user

Saturday, June 18, 2005

Meeting Fri: 6/17/05

:presented Marketnet approach to find equilibrium point. ppt

:the result from the simulation. ppt

Thursday, June 16, 2005

Simulator Log

HAPPY BIRTH DAY FOR MY PLATFORM and AGENT.

THEY ARE BREATHING TODAY.





Below is the simulator log:

Date 6/15/05
Agent: Energy Exchange Algorithm
0) Energy Algorithm class is implemented for Agent
1) Agent is implemented Energy Algorithm (as object).
It keeps record of intake energy from doSevice behavior.
2) Agent pay energy (check and pay) in preDecision behavior


Platform: Energy Exchange Algorithm
0) Platform has EnergyExchangePF class
1) It keeps record of intake agent when the agent give energy to platfrom
2) It pays energy in preDecision behavior
3) doEvaporate behavior check the healthy level for platform








Date 6/1/05
Energy Exchage Algorithm
Infomation
1) mainSimFrame Cycle_num is the main number in cylcle
we cannot use this number for keeping the time






Date 2/9/05
Plan Next:
1) CE give energy to Platform
2) CE
Show the result after platform as biological entities:



Date 2/7/05
Healthy level completed
- platform use 10 unit of each HL
- Agent use 1 unit of each HL
- Platform will check the number of CE when it respiration and update



Date 2/5/05
Done:
Platform work on host use 10 healthy level each
Agent work on host use 1 healthy level each
This algorithm is added in BionetPlatform.java
* BionetPlatform --- control platform number, CE number .. . .
Platform <-- BehavPlatform control platform as biological entities
Effect the replicationThreshold of Platform control the number of platform
2 the healthyLevel of platform are negative
3 Agent did not give any energy to platform when it work on
4 agent did not pay healthylevel back to platform when it maigrated






Date 2/4/05
Added Node file field CPU MEM NETWORK
Changed FileParser to read the CPU,MEM,NETOWRK field from node file
Added constructor to node( ) class
Added BionetPlatform in setPF to init all HL
setPF set Healthylevel of platform =node minus 10 for platform cost
Added Platform set HL
Changed GUI control panel to show HL
ControlPanel was changed to check Platform enable
2/5/05 Platform has Healthy level





Date 1/30/05
Added Gui Platform as Biological Entities
control panel
and change in simulator

Date 01/24/2005
GUI
COLOR
requestNum ==0
Node Provide Service > 0 : Blue (CE in Service)
CeNum=0, : white (No CE)
CeNum>0 and provide service =0 : Yellow (CE Idle)
REquestNum>0 : Red
if node has last request >0 red circle
platform green
16:25:08, Improve some platform gui



Date 1/23/05
To show the result of Platforms
+ Changed About Dialoge
+ Added PlatformPanel by copying ResultPanel
+ Changed CE to PF

+ Added parseOutputPFFile in FileParser.java
+ Added ResultPF.java to keep the tempplate of Platform

+ Added creating pf.xls in OuputDatamanger.java

+ Changed OutputData in Platform to add information to the file

+ Added code in endsimulation in BionetPlatform

Friday, June 10, 2005

My Stuffs

These are the files I have about the symbiotic network

Summary Slides (85 pages)
- Symbiotic Architecture for Adaptive Network PPT

Drafted Slides (for 15 minutes talk)
- visitor(Prof) present


Poster
- Alumni Party (Big Poster in PPT)


Basic Idea about forecast and represent data
PPT file 2 types of Moving Average Algorithm
- SMA
- EWMA (+ why it is exponential0 )

ภาษาไทย
- ก่อนมันจะหายใจ

Tuesday, June 07, 2005

Today Update

I have nothing to present about my research progress today.
The agents and platforms have not taken the first breath yet.
It is very late and far behind from my schedule.
Everything will be finished by this weekend.
I will work harder and harder. Please Wait.

Today Meeting:
CN presented about the self-assembly and related work.
He was killed again.

My turn, I presented EWMA why it is exponential. And also,
the marketnet from columbia university.
Nearly be dead, but I was not.


The proposal idea, the good idea was poped up.
We can implement bio-net to animal observer system.
The agent works as cache.
The goals are network traffic and database space reduction.
The local of agents has been proposed.

We used Chiangmai Elephant as a case study.




Task:
1) My CV due this night
2) REview Paper x3 due Tomorrow
3) Paper due June 20
4) Simulation Result due this weekend
5) Prepare for Next Meeting this Friday

Tuesday, May 31, 2005

Stuffs in Our Lap Space

**** Our Stuffs ****

There are 8 monitors in our lab space
( 19" (Paskorn's) ,+ new 19", + 15"x6)

There are three PCs (Paskorn's 3.0 HyperThread 1GRam, Linux Server, and Chonho's PC)

The new PC is comming for Madhu.

1 samsung laser printer

1 wireless D-link router (G)

1 4 port linksys switch

1 4 port 10M HUB (the old one).

1 Refrigerator

1 Microwave


The stuffs' manuals are in the Chonho's drawer.


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




Meeting: Research Progress::

Now, we have 2 meetings a week:

Chonho presented about self-assembly and origami langauge.
He will present more related works this Friday.

Paskorn presented about market net for sensor network (Done).
He will present related work in market net and his research progress.

Paskorn is working on Energy Exchange Algorithm. The GridNets paper is dued this June 15th.


Task:

For grant proposal: How to apply bio-net? (Due Friday)
Review 4 papers (Due Friday)

Sunday, May 15, 2005

Some memo notes

Something that I should write it down before I couldn't remember.

1) I read MALLET paper for review.
The mallet is Multi – Agent Logic Language for Encoding Teamwork. The authors are attempting to create the new langauge to implement program agent as a team set.

For instance, in the fire fighter team, there are many agents in diferent function but they have the same global goal. This problem can be implemented by MALLET.

The syntax and semantics are hard to understand for me, cause , again, I am not a guy who is good in math. :)

The point is what happens if we implement agent and platform in my model follow this concept.

My model there are agent to serve the user and platform to find the better host to be installed. Both of them have the same finally goal, survival. To let them survive, agents have to serve the user. We can imaging the serving request from user as a web page request and each agent contains web page. When the agent get the request it get the energy (assume set value). Energy drive them to living and behave. Agent and platform live together as in symbiotic concept.

Come back to the point, what happens if we implement mallet. Yes, it should be a great idea. Now I create all symbiotic behaviors up(hard coding). But, MALLET can invoke these behaviors automatically. The new language to describe the mechanism of network adaptation should be a good topic for doing research.

Yes, I have to study more how to implement this concept in my model.


2) Another thing, I just presented poster about the symbiotic network in computer science UMB last Thu.

Next step, we are going to Orlando, Florida in this July to present this small idea.

The world, see you there.

Energy Exchange Algorithm

Long time, I've not updated this blog because there was no surprising progress and I was very busy.

This semester is going to kill me. Studying Compiler and OOD together is so real hard. For me, I loss my way in the wrong track which I miss-chosed.

However, the symbiotic networking has some progress. Since last update, I came back to basic questions: how a life survives in the real environment, how energy relate to the living.

Of course, I read many biology papers, which I don't like them much since I was in high school age. I learnt how the salmon-fish use energy for reproduction, how animals transfer the food to energy and make them still living :(metabolism). I knem that there are some theory in biology as 10% rule: energy transfer, in food chain, in each level only 10% that it takes.(Pyramid of Energy)

First I attempted to apply hibernate algorithm, but I couldn't. We don't know the user predicted request period, of course we couldnot forecast it either. Some theories came up! morkov chain, queing theory but again I am a guy who is not good at mathematics.

Finally, the easy basic idea was poped up in my head. Just find the average, maximum and minimum of interval time (we have to change frequency domain to time domain). And then let agent, as biological entity, losses its energy based on these values as respiration rate.

The result showed surprisingly.

Sure, now, I have to implement this concept to the bio-net simulator.

Do you know what: this is far from the symbitic behavior. Now they, living agents, just start breathing.

Anyhow, So far so good.

Wednesday, February 09, 2005

2/9/05

Happy Chinese New Year Day!

Simulator:
SymbioticSphere
Now, the platform is alive. It can replicate, dead, and maintain healthy level.
The node (or host) in the network also have resource availability -- (CPU,MEM,Bandwidth)

They, platforms and agents, live together in the SymbioticSphere as the living
entities that require each other to survive.

Next Step:
I will implement symbiotic behaviors.
Of course, we have to find to evaluate the simulator somehow:


Related Work:
We are studying cell cpu.

Thursday, January 27, 2005

Progress 1/27/05

Done:
+ update drafted paper
The paper has been revised in section 1-4.
- The symbiotic behavior names have been changed (fig 3,4)
- Healthy level table has been added in the platform attributes section
- I rewrite the section 4: symbiotic network.
- I also added the related work section in the drafted paper (section 6)
* The simulation part need to be added soon. I am waiting for simulator
programming.

Monday, January 24, 2005

Progress 17-23 Jan 05

Done:
+ Model platforms as biological entities
- Platforms are deployed to nodes (hosts) by using deploy file (file.pf). The pf file also contains platform attributes. (initiate energy, node)
- Platforms can replicate, death
- The graphic result when simulator working is showed.
- The result of platforms after simulation is showed. (#platform alive/death, generation, energy, #replication, birth time, life time, etc. )

Task:
I will do:
+ Simulator Programming
- Add healthy level to host (node)
- Show some result that can be used in the sci05 paper.
+ Yan's document comment
+ Paper sci05
- revise section 4 (Symbiotic network)
- layout simulation section
+ Prepare for related work present on Friday

Sunday, January 23, 2005

Progress 1/23/05

DOne: 03.00am
Simulation:
+ Platform V0.5
(with No Error)
Platform can deploy, replicate, dorespiration and death




Task (Todo Next)
+ Add Platform panel to show Platform output + graph
+ Add Result Panel to show expected result
+ Add healthy level and healthy level table
+ Add Symbiotic behavior
+ checked box to set options


Date 1/23/05
Done: Show the result of Platforms

Simulator Modified Lists:
+ Changed About Dialoge to show version 4.5 Symbisis NetSphere ;)
+ Added PlatformPanel by copying ResultPanel
+ Changed CE to PF

+ Added parseOutputPFFile in FileParser.java
+ Added ResultPF.java to keep the tempplate of Platform

+ Added creating pf.xls in OuputDatamanger.java

+ Changed OutputData in Platform to add information to the file

+ Added code in endsimulation in BionetPlatform

Progress 1/22/05

Done:
Simulation
Platform take a first breath()
Platform can replicate to other node.
Platform can be dead

Gui update to Platform

Saturday, January 22, 2005

Progress 1/21/05

Done:
Simulation
- Initial Platform to simulator
Deploy platform with pf file
Platform GUI on network panel
Platform gains energy from agent
- Agent needs platform to work on. (by changing ServiceCE.java migration behaviors)

Added Files
bioarchitecture.entitties
- Platform.java //Main platform
- BehavePF.java //contains platform's behaviors / extends Platform
- PlatformClass.java
- PlatformClassList.java

Changed file
bioarchitecture.gui
- Control Panel.java // Added platform GUI to access .pl file
- TopologyDrawPanel.java //added details of platform in the node

bioatchitecture.nodes
- BionetPlatform.java //Added platform details numPF, getPF()

bioatchitecture.simulator
- DeployData.java
- FileParser.java

Tuesday, January 18, 2005

1/18/05

Done:
- Register for CS696
- Get CS620 Exam back A

Simulator Done:
- Add Platform to the first page,
I will add platform config to the simulator for setting
Initiate platform, energy, which nodes


Monday, January 17, 2005

My progress 9-16 Jan

Done
1. Commented Yan's document
2. Submitted first drafted of related work
3. Revised sci05 paper section 1-3.
4. Submitted simulation plan revise#35. Simulation program
+ added static node 4x4
I am doing:
1. Simulation programming
2. revise sci05 paper section 4
3. lay out simulation section in the paper
4. Prepare to present next Monday "paper eading presentation"
5. Add comment to Yan's Document

Saturday, January 15, 2005

My Progress 1/15/05

Done:
- Powerpoint structure of agent and platform
- revise sci05 paper (done in agent part)

Thursday, January 13, 2005

My Progress 1/13/05

Done:
1. Review and correct sci05 paper. (4.00am)
- Add figure 3,4 agent and platform structure.
- Change word "regular behaviors" to "behaviors"
- Change word "symbitic behaviors" to "symbiotic mechanisms"
- For now, each agent consists of four parts: attributes, body, behaviors, and symbiotic mechanisms.
- Add healthy table part with example table
- Add some paragraph to describe symbiotic mechanisms in high level concept.
- Lay out the related work but NOT FINISH

http://dssg.cs.umb.edu/~paskorn/Bionet/My_Paper/Sci05/sci05_symbioticnetwork_PCR2.doc

Task:
Simulation!!!

PS.
On Thursday, I will not be available after 3pm
http://calendar.yahoo.com/paskorn

Wednesday, January 12, 2005

My Progress 1/12/05

Today:
Kuitar departed to Bkk at 11.00 AM.
I am alone and have nothing to do but hard work.

Done: (7.00pm)

  1. Read related paper

“Dealing with ghosts managing the user experience of autonomic
computing”, D.M. Russell, R.P. Maglio, R. Dordick, C. Neti, IBM
Systems Journal 42, No. 1, 177-188, 2003

“Enabling Autonomic Behavior in Systems Software with Hot
Swapping,” J. Appavoo, K. Hui, C. A. N. Soules, R. W. Wisniewski,
D. M. Da Silva, O. Krieger, M. A. Auslander, D. J. Edelsohn, B.
Gamsa, G. R. Ganger, P. McKenney, M. Ostrowski, B. Rosenburg,
M. Stumm, and J. Xenidis, IBM Systems Journal 42, No. 1, 60-76,
2003.

  • I 've found the web site http://www.thefutureofcomputing.org/
    The future of Computing, pervasive, autonomic, grid computing.
    78-page book. It is easy to read for beginner.
  • I listed the related papers that I've read. They are about 10s in my list right now.

2. Start Writing lay out of related work.

3. Simulation (Done at 11pm)
Add simulator to Eclipse, run test,
Change button picture icon
Read the manual.

Task:
- Add network topology to 4x4 Static Done at 11.45pm
- Add one more type of agent (to be platform in the future)

Memo:
My First night in the campus, no where to go.


Monday, January 10, 2005

My Progress 1/10/05

Done
- Read realated paper about pervasive network +2


- Add more in reveiw paper power point about
- Autonomic computing


Task
Search and read these papers, about resource management in dynamic network

- C. Ragusa, A. Liotta, and G. Pavlou. Dynamic resource management for mobile services. In The 5th International Workshop on Mobile Agents for Telecommunication Applications (MATA’03), 2003.
- L. Ismail and D. Hagimont. A performance evaluation of the mobile agent paradigm. In the Conference on Object-Oriented Programming, Systems, Languages, and Applications, pages 306–313, 1999.
- Robert S. Gray et al. Mobile-agent versus client/server performance: Scalability in an information-retrieval task. In The 5th International Conference on Mobile Agents(MA2001), pages 229–243, 2001.
- Zakaria Maamar, Quan Z. Sheng, and Boualem Benatallah. On composite web services provisioning in an environment of fixed and mobile computing resources. Information Technology and Management, , Kulwar Academic Publishers,, 5(3), 2004.


Write Paper
- Revise sci05 Paper
  1. Add "healthy table maintain"
  2. Write simulation plan
  3. Write lay out related work

- Program

- Add one type of static network


Saturday, January 08, 2005

My Progress 1/9/05

Done
  • Reading paper and revised version 1
  • Reading realated papers
  • - Autonomic Computing x 3 Papers
  • Study some part of Simulation
  • Submit comment for simulation document (Yan's document)
Admin work
  • Got fist payment
  • Applied for TA (cs446: internetworking)

Nest Task (not done)
  • Simulation document comment on the file
  • Update PPT paper review
  • Revise main document (sci005 document)
  • + Add realate work
  • + Simulation Plan
  • Write Code
Plan
+ simulation code
I'll add one more agent call platform so we will have regular agent and platform agent.
Then, I will write some code to let regular agent require platform agent to work on the host first.
The platform agent works as platform entity in my research. Therefore, we can simulate and get some result from this idea.
The problems, this is not totally true as we plan in the research. But it save the times and it is good step to be samiliar with the Java programming. The next step we can change the place of platform service code to be in the platform agent.

+ Document
I have to write related work and simulation as soon as possible. The first version should be finished before this Tuesday!