490 likes | 591 Views
Le-Chi ’ s log. Since 2004/01/05. Weekly Plan 04/11/01. The last plan QoS : pre-benchmark by avalanche Webfd : NAT in iptables Computer Network : midterm The plan for this week QoS : pre-benchmark by avalanche QoS : squid & apache proxy evaluation
E N D
Le-Chi’s log • Since 2004/01/05
Weekly Plan 04/11/01 • The last plan • QoS : pre-benchmark by avalanche • Webfd : NAT in iptables • Computer Network : midterm • The plan for this week • QoS : pre-benchmark by avalanche • QoS : squid & apache proxy evaluation • Computer Architecture : midterm preparing • Last week, we discussed the possibility to implement my thesis algorithm on open-sourced proxy package instead of webfd. And This week, I plan to survey them to see if they are appropriate target to have my QoS algorithm. Hope I can make progress on benchmark by avalanche.
Weekly Plan 04/10/26 • The last plan • QoS : patent database • QoS : pre-benchmark by avalanche • Webfd : NAT in iptables • Computer Network : hw2 & midterm preparing • The plan for this week • QoS : pre-benchmark by avalanche • Webfd : NAT in iptables • Computer Network : midterm • I searched the US & TW patent database, couldn’t found anyone exactly based on the same algorithm as ours. Last week, I installed webfd in NetBSD, both the multi-process and single-process version, for test and realizing why NAT is needed. And, I still can’t resolve the problem with avalanche.
Weekly Plan 04/10/19 • The last plan • QoS : pre-benchmark by flashget • Linux Wall : hostap GUI • Computer Architecture : SimpleScaler lab • The plan for this week • QoS : patent database • QoS : pre-benchmark by avalanche • Webfd : NAT in iptables • Computer Network : hw2 & midterm preparing • Last week , I try to do my research pre-benchmark by avalanche. It has complicated setting and many steps before starting any testing. I think I will spend much time getting familiar with it. Webfd is implemented for NetBSD originally, I need to find out the NAT transformation in iptables for my first step to porting webfd to Linux.
Weekly Plan 04/10/11 • The last plan • QoS : verification and recording to extracting URL lists for pre-benchmark • Linux Wall : hostap policy & merging into wall image • Computer Architecture : hw2 • The plan for this week • QoS : pre-benchmark by flashget • Linux Wall : hostap GUI • Computer Architecture : SimpleScaler lab • I extracted the useful URLs from the squid log mess last week. From discussion, we think we can do the benchmark work by flashget instead of avalanche. I plan to finish it in the end of this week. As for Linux wall, the code base in CVS is not stable yet, so it needs more effort to merge hostap into linux wall code base.
Weekly Plan 04/10/04 • The last plan • QoS : proxy log analysis & paper reading • Computer Architecture : paper reading • Network Security : how to find all the IP in used in your local LAN • The plan for this week • QoS : verification and recording to extracting URL lists for pre-benchmark • Linux Wall : hostap policy & merging into wall image • Computer Architecture : hw2 • I started to extract useful URLs from the proxy log we got last week. It costs much time to filter the un-suitable URLs for the mess. Hope I can finish the filtering job, and start to benchmark to see if there is any question exists in my algorithm.
Weekly Plan 04/09/29 • The last plan • Wall : hostap • QoS : pre-testing for network bandwidth usage • Lab website maintain : tagging lab research publications with SCI or EI • The plan for this week • QoS : proxy log analysis & paper reading • Computer Architecture : paper reading • Network Security : how to find all the IP in used in your local LAN • In order to get realistic proxy log, I tried several ways last week. Fortunately, we got tremendous proxy log from certain unknown company finally. Also, I searched all the research publications in the SCI and EI database, and tagged them if they were hit. The six-month plan is in the next page.
Weekly Plan 04/09/20 • The last plan • Wall : hostap • QoS : layer4~7 workshop • Taking courses • The plan for this week • Wall : hostap • QoS : pre-testing for network bandwidth usage • Lab website maintain : tagging lab research publications with SCI or EI • I tried to setup a linux wall machine for build hostap last week, because hostap uses modules that no other packages use too. It’s harder to achieve it than other packages. Hope I can make it work in this week.
Weekly Plan 04/09/14 • The last plan • Wall : static route & policy route GUI • QoS : paper surveying • The plan for this week • Wall : hostap • QoS : layer4~7 workshop • Taking courses • Summer vacation is finished. We have an almost-perfect lab trip in the last week of summer. Although every day is rainy day, it’s still funny to race with the whether. Thank every labmate for giving me a great summer ending.
Weekly Plan 04/08/31 • The last plan • Wall : misc & route GUI • QoS research : proposal preparing • The plan for this week • Wall : static route & policy route GUI • QoS : paper surveying • Last Friday, I have a high ranking for my thesis proposal. It’s because weafon’s directions. However, it lacks support of related works. So I must complement this part in the next step. As for Wall GUI, because the different architecture of BSD and Linux, it needs some time to adjust the GUI skeleton and make it work.
Weekly Plan 04/08/23 • The last plan • Wall : misc GUI • Wall demo in Taipei • QoS research : flow chart and implementation position • The plan for this week • Wall : misc & route GUI • QoS research : proposal preparing • After researching all this summer, we have made a model for my thesis objective. But it still lacks a strong and accurate title. Moreover, we have to confirm it’s scale by experiment in advance. As to Wall GUI, I have finish a part of misc modifications. It’s not interesting, but just doing it.
Weekly Plan 04/08/17 • The last plan • Wall : GUI for route function • QoS research : proposing new RPC and refining new RTC architecture • The plan for this week • Wall : misc GUI • Wall demo in Taipei • QoS research : flow chart and implementation position • We proposed a new method for “request scheduling for differentiated QoS at access gateway” last week. The idea of RTC is alike best-effort. And the concept of RPC is a variety of WFQ. The workload of new RR is less and simple.
Weekly Plan 04/08/10 • The last plan • Wall : discussing and distributing the workload of GUI • QoS research : looking for research issue from Ou-Yang’s subject • The plan for this week • Wall : GUI for route function • QoS research : proposing new RPC and refining new RTC architecture • Last week, we propose a new idea and rough RTC architecture instead of Ou-Yang’s original method in differentiated QoS request scheduling. We don’t know if it’s a good method for this issue, so we still need to discuss and refine. As for wall GUI, my job is implementing the user interface of route and hostAP.
Weekly Plan 04/08/02 • The last plan • Wall : resolving pcmcia-cs packages problem • QoS research : surveying related paper, discussing with Ph. D mentors • The plan for this week • Wall : discussing and distributing the workload of GUI • QoS research : looking for research issue from Ou-Yang’s subject • I resolve pcmcia-cs packages problem last week. I modify the kernel configuration, and then we can successfully “make” it. Today, I have my wisdom tooth pulled out. The swelling cut sheds blood all day, and I can’t eat food joyfully. So bad a experience.
Weekly Plan 04/07/26 • The plan for this week • Wall : resolving pcmcia-cs packages problem • QoS research : surveying related paper, discussing with Ph. D mentors • At last I accomplish all the assignments of ISAAC. It is a huge work. All of them are 7 homework and 1 term project, and I mush complete them during 2 weeks!! Oops! Taking 3 hands-on oriented courses in a semester is too busy to take care of all tasks well. Now, I get rid of ISAAC, I’ll try to come up with the research scheduled progress.
Weekly Plan 04/06/13 • The last plan • Wall : leaf buildroot • EOS final project • The plan for this week • EOS final project : Access Point on ARM • It’s almost the end of the semester, And I need to finish 1 term project and 3 network experiments in the 2 weeks. Taking 3 hands-on courses is so terrible a decision. The term project of EOS is very hard to achieve. I’ll try my best to finish it before the end of this week. The late of last week, all the MS 2 people have passed their oral thesis exam. All of them have a great result for the 2- year effort. I’m so proud of them.
Weekly Plan 04/06/07 • The last plan • EOS presentation and report : linux filesystem – jffs • Wall packages functionality testing • The plan for this week • Wall : leaf buildroot • EOS final project • We find a new solution for wall these days. It’s leaf buildroot. We have tried several solutions this semester. Hope it’s the final one. Finally I finished the EOS homework 2. The load of this course is very heavy. Actually, all the course I take are all heavy load. That’s why I always stay up all night. Thanks God it’s almost the end of this semester.
Weekly Plan 04/05/31 • The last plan • EOS homework : linux filesystem – jffs • ISSAC homework • The plan for this week • EOS presentation and report : linux filesystem – jffs • Wall packages functionality testing • Last week is a busy week. The early days in last week, I spend all my time in tracing linux jffs file system code. It’s a difficult job. And the later days in last week, All wall team members fight for the contest in weekend. Although we fail in the contest, we’ve try our best. I’d like to thanks for ycchang’s parent. The provide us the housing, and treat us a dinner, that saves us a lot of money and time for surveying the information of hotel.
Weekly Plan 04/05/25 • The last plan • Adding DansGuardian (and maybe rrlogind) to buildroot • EOS homework : arm-linux filesystem • The plan for this week • EOS homework : linux filesystem – jffs • ISSAC homework • Finally, I’ve added DansGuardian to buildroot successfully. Because DansGuardian is written by c++, it must use not uclibc but g++ library. As for EOS homework 2, I will trace the linux kernel code for jffs file system. It’s a big job for me. It involve many components. Besides VFS, inode, dcache, I also need to know how system call works; what is spinlock…etc. To understand it fully, it costs me lots of time undoultedly. However, time is not enough for me now.
Weekly Plan 04/05/17 • The last plan • Computer Network Experiments : VPN and Cisco router • Isaac homework : buffer overflow • Adding Squid and DansGuardian to buildroot • The plan for this week • Adding DansGuardian (and maybe rrlogind) to buildroot • EOS homework : arm-linux filesystem • I’ve added Squid to buildroot with some patch. And I’m stilling fighting for DansGuardian until today. Hope I can finish it tomorrow. I spent most time on Computer Network Experiments homework last week. It’s a good course, because we can gain hands-on skills and in the same time, we must answer some question about the tech aspect. The only flaw is most experiments are out of date.
Weekly Plan 04/05/10 • The last plan • Adding packages to buildroot • Computer Network Experiments : VPN • Isaac homework : buffer overflow • The plan for this week • Computer Network Experiments : VPN and Cisco router • Isaac homework : buffer overflow • Adding Squid and DansGuardian to buildroot • I added net-snmp and pure-ftpd to buildroot last week. And this week, I hope I can added Squid and DansGuardian to buildroot successfully. How to make them use the right library module is a troublesome issue. As for VPN experiment, it’s a little time-consuming. It costs us more time than we estimate. Any way, we will close it this week.
Weekly Plan 04/05/03 • The last plan • Finishing building snapgear • EOS homework : term project SDS • Computer Network Experiments homework • The plan for this week • Adding packages to buildroot • Computer Network Experiments : VPN • Isaac homework : buffer overflow • After working on snapgear for more than 1 week, we discover it’s very difficult to add new packages. Even though making the packages it has originally, it’s not easy to make it. Therefore, we will get back to buildroot. Moreover, two homework are waiting to finished.
Weekly Plan 04/04/26 • The last plan • Testing the packages added into the buildroot and modify makefile • EOS homework : bootloader report • The plan for this week • Finishing building snapgear • EOS homework : term project SDS • Computer Network Experiments homework • There are some problems when I try to build snapgear. I plan to build it sucsessfullly in this week. Last week, when I start to work on computer network experiments, I notice many of them have gone out of date. We must modify experiment content so that we can keep working on it. Something awful is my front teeth is broken today. I need to make a cover on it.
Weekly Plan 04/04/19 • The last plan • adding the packages into the buildroot • EOS homework : bootloader • The plan for this week • Testing the packages added into the buildroot and modify makefile • EOS homework : bootloader report • I installed all the packages last week. I’ll verify their functionality in these two days. After that, I’ll modify their makefile to compliant the wall Linux's in the following week. This semester, we have much load in the course in Embedded OS. We spend much time in hands-on with the evaluation board, with the OS, and with the development tools. But also we learn much from it.
Weekly Plan 04/04/12 • The last plan • Building wall packages in Linux through virtual machine • EOS homework : bootloader • The plan for this week • adding the packages into the buildroot • EOS homework : bootloader • Last week, I get the testing wall linux environment through CVS and build it. After that, I try to add the packages into it. For the pourpose, I must modify the packages' Makefile to compliant the wall linux's, and it's in progress. About the EOS homework : bootloader, finally I resolve the problem that I can't download image from CF card. It's because the manual doesn't depict it clearly. That’s a problem shouldn’t be a problem!
Weekly Plan 04/04/05 • The last plan • Wall package surveys • EOS homework : bootloader • The plan for this week • Building wall packages in Linux through virtual machine • EOS homework : bootloader • Last week, I spent my most time on being familiar with the evaluation board I got. I tried to download image through CF card – the only way it can accept new image from outside. I followed all the instructions from the document, but it didn’t work. And the TA can’t give me any useful answer or guide, cause’ they didn’t work on it before. Hope so much I can resolve the problem this week.
Weekly Plan 04/03/29 • The last plan • Wall package surveys • EOS homework • The plan for this week • Wall package surveys • EOS homework : bootloader • About the wall surveys for Linux, I’m responsible for the 5 packages : ftpd, DansGuardian, Squid, SNMP, and rrlogind. All of them are capable of Linux or can find substitutes. Besides, I also need to trace the code for macf, kva, and webfd. I’ll finish them in two days.
Weekly Plan 04/03/22 • The last plan • 802.1x + radius for HostAP function in Linux • DansGuardian benchmark • The plan for this week • Wall package surveys • EOS homework • Last week, I experiment the port access authentication of HostAP by package “hostapd” and “WinRadius” running in Windows OS. Next time, I will experiment it with FreeRadius server. Unfortunately, I got the 落枕 several days ago, it’s so bad an experience.
Weekly Plan 04/03/15 • The last plan • HostAP function in Linux • DansGuardian benchmark • The plan for this week • 802.1x + radius for HostAP function in Linux • DansGuardian benchmark • I setup a hostAP in linux last week, basic function can work now. And I plan to build 802.1x and radius function this week. As for DansGuardian benchmark, we have the performance results for the combinations by if enabling cache and early decision. The next 2 slides have the result charts showed.
DansGuardian performance for early decision and cache • Request per second
DansGuardian performance for early decision and cache • Throughput
Weekly Plan 04/03/08 • The last plan • HostAP function • DansGuardian benchmark • The plan for this week • HostAP function in Linux • DansGuardian benchmark • Since we will port WALL to Linux in the near future, I plan to find the solution for building hostAP in linux, and setup it in this 2 weeks. Even though we will suspend our small function developed in NetBSD, and it’s not completed yet, we still get some experiences from the past weeks. Any experience is a gain for me.
Weekly Plan 04/03/01 • The last plan • DansGuardian benchmark for Chinese content • Setting up NetBSD-current and Open1x for hostAP function • Prepare documents about DansGuardian for contest. • The plan for this week • HostAP function • DansGuardian benchmark • I tested the accuracy of DansGuardian works on Chinese content. The result is showed in the following slide. I built NetBSD-current kernel to experiment on hostAP function. But the pcmcia card can’t be found until now. I’ll keep working on it this week and make progress.
Weekly Plan 04/02/23 • The last plan • Paper presentation • Taking courses • Setting up NetBSD-current and Open1x for hostAP function • The plan for this week • DansGuardian benchmark for Chinese content • Setting up NetBSD-current and Open1x for hostAP function • Prepare documents about DansGuardian for contest. • This semester, I’ll take 3 courses. All of them are focus on hands on works. Undoubtedly, it will be a very busy semester for me. The progress of hostAP in wall is delayed, hope I can catch up during this week. I know there are much can be and must be improved about my presentation quality. I’ll make it better gradually, even though it’s slightly growing.
Weekly Plan 04/02/16 • The last plan • Paper reading • DansGuardian benchmark • The plan for this week • Paper presentation • Taking courses • Setting up NetBSD-current and Open1x for hostAP function • Last week, I re-tested accuracy of DansGuardian because the unconvincing result from last testing. The result is showed in the following slide. As for HostAP function, NetBSD just supports it in the current version, and it doesn’t support 802.1x. The only one 802.1x package can be run in BSD platform that I found is open1x. However, it’s still not a mature package now. So the thing I should do in this phrase is trying to set up this solution.
Weekly Plan 04/02/09 • The last plan • Paper reading • DansGuardian benchmark • The plan for this week • Paper reading • DansGuardian benchmark • Last week, I had a test to find out why the performance of URL-based filter is worse than content -based filter in our previous testing result. We suppose that’s because the patterns of URL-based filter are much more than content-baseds’. So we made them have the same number of patterns and tested again. The testing result are depicted in the flowing 2 pages. As for HostAP function, most packages supply to Linux. FreeBSD also supports it by kernel. NetBSD just supports it in the current version.
Comparison between URL and Content • request / second
Comparison between URL and Content (cont’) • throuthput
Weekly Plan 04/02/02 • The last plan • Computer Network mini project • Wall demo • The plan for this week • Paper reading • DansGuardian benchmark • This week, I have 2 task about DansGuardian benchmark. The first is re-test its accuracy because the unconvincing result from last testing. And the second is try to figure out why the performance of URL-based filter is worse than keyword-based filter. Besides, I’ll reading the system research paper for the first time.
Weekly Plan 04/01/12 • The plan for last week • DansGuardian performance benchmark for small web page about 5KB • preparing 3 final terms • The plan for this week • Computer Network mini project • Wall demo • Last week, I discussed how to test CPU loading of DansGuardian with HFH, the result is showed in next page. Besides, I re-test a questionable case of performance. Last week is very busy, I think all MS1 labmates are very tired, especially in end of last week. We stayed all night to work for wireless project. So unforgettable an experience!
Weekly Plan 04/01/05 • The plan for last week • DansGuardian performance benchmark for small web page about 5KB • preparing 3 final terms • The plan for this week • 3 final terms • Wireless ad hoc routing project • DansGuardian re-testing for some questionable result. • Last week, I did DansGuardian performance benchmark for 5KB white page, but it seems some of them are questionable, I’ll redo it before the end of this week. About observing the CPU loading when DansGuardian’s running, I’ll try to figure out how to make a meaningful testing way, maybe by command “top”. There are 3 final terms this week, and no enough time to prepare them. Hope I won’t get low scores.
Benchmark for DansGuardian • 5KB white page, 8 clients • requests / second