20150812 SLAC SEECS and UNIMAS Meeting

Time & date 

Next meeting:  Wednesday August  12th 2015 9:00pm Pacific Standard Time, Thursday August 13th 2015  9:00am Pakistan time, Thursday August 13th 2015 noon Malaysian time, Thursday  Thursday August 13th , 2015 02:00am Rio Standard Time.  

Coordinates of team members:

See: http://pinger.unimas.my/pinger/contact.php

Attendees

Invitees:

Hassaan Khaliq, Kashif, Raja,  Samad Riaz (SEECS); Johari+, Nara, Adnan Khan? (UNIMAS); Abdullah, Badrul, Anjum+, Ridzuan, Ibrahim? (UM); Hanan, Saqib- (UTM); Adib-, Fatima+ (UUM); Fizi Jalil? (MYREN);  Thiago+, Les+, Bebo- (SLAC)

+ Confirmed attendance

- Responded but  Unable to attend: 

? Individual emails sent

Actual attendees:

Saqib, Thiago, Les. Johari unable to join due to security blocking Skype. Johari  joined later after Saqib left.  We called Fatima, Anjum, Ibrahim several times but with no success. Anjum was involved in a traffic accident and was uabel to attend.

Administration

• Membership of pinger-my in https://groups.google.com

• We are working with Dr Zaidi the new head of SEECS to investigate resources need to continue support of PingER in Pakistan. We communicated last weekend.

• NETAPPS2015
  

  • The final submission date is now Aug 25th August.

  • Anjum submitted a paper to NETAPPS2015 "Adaptive Geolocation of Internet Hosts"

  • Les, Thiago, Johari, Bebo and Topher White submitted a paper on "Worldwide Internet Performance Measurements Using Lightweight Measurement Platforms"

  • Saqib submiited a paper "PingER Malaysia-Internet Performance Measuring Project: A Case Study". This paper needs reformatting, Saqib is working on it. It also needs reviewing. Les will look at.

  • Thiago and the Brazilian team submitted a paper on "Applying Data Warehousing and Big Data Techniques to Analyze Internet Performance'. Adib has kindly come up with a way in which the authors are unlikely to be able to attend, so somebody else will present.

• Saqib's abstract (basically an abstract of his NETAPPS2015 paper, it is not for publication) submission  "PingER and Internet Performance in Malaysia" was accepted for presentation in APAN-40 (https://www-lk.apan.net/meetings/KualaLumpur2015/schedule.php). APAN-40 will be held in collaboration with UM, UPM and MYREN from 10-14 August 2015 in Research Management & Innovation Complex University of Malaya. Saqib was unable to attend since his visa did not arrive in time.

• Congratulations to Saqib he received the fellowship from APAN-40 regarding traveling and accommodation in Malaysia. Further, he will receive the airfare from his university in Pakistan.  However, his visa did not arrive in time so he was unable to take advantage of the offer.

• Johari held a colocated PingER/BigData workshop on August 3rd the day before the  CITA 2015 (see http://www.cita.my/ an International Conference 4th - 6th August 2015, on transforming Big Data into Knowledge. Bebo made a presentation. Bebo used PingER as a case study on what you can do with PingER and how to access. An issue raised was how do they get hold of the data. There was a discussion between Bebo and Johari. Users can access the  data via FTP, this repository can easily be updated on a regular basis (Les has the code). However, Bebo and Johari would like to make the data available  via the web observatory hence making it more public.  Renan collected data from FTP and converts data in a star schema model. Thiago converted to Hadoop and Impala, currently stuck in a security problem, working on Kerberos to try and set up security, may require a bastion host to go through to access the data. Thiago still unsure if public for a while. Web observatory user downloads the data and then uses  (i.e. do not query directly). Web interface uses triplets. There is a impala shell that enables SQL access. 

• Bebo is back from a trip to UUM & UNIMAS.

• Anjum believes the TULIP Geolocation application  can be improved significantly. At least there are few ideas that we can try. For this, either a group of undergraduate students or an active masters student is required. The resultant work can easily be the thesis of masters level. Who is interested? 

  • Saqib is going to see if he has a student interested. He will contact Anjum to learn more.

  • Johari will contact Anjum to learn more of the requirements.

UFRJ

 UUM

Unable to gather data from pinger.uum.edu.my since July 12th, 2015.

Fatima has installed hadoop on all three machines. One will stand as the host machine from which the remote machines will be controlled. She is currently trying out some MapReduce examples and Hive installation.  

They have managed to get a copy of PingER data for Fatima research from UM friends, Thanks to Dr. Anjum and Mr. Ibrahim for their help and support. Update

UM

We have successfully gathered data from pinger.fsktm.um.edu.my since July 9th 2015. Also www.um.edu.my is not pingable. Anjum reports that it is old and needs replacing.

Ibrahim had downloaded PingER in Zip files format, however, when he stored them in the Hadoop distributed file system (HDFS) and try to process them, the file got corrupted, so he had to extract the file, but one file zip has more than 10000 zip files with small size.  So he is trying to create a mapreduce job which can accept zip format, that will save lot of his time. Currently mapreduce can only read from files like .txt, and any doc file format or database. He will have meeting with Dr. Anjum on 11 of june asking for advice and seeking of how we can work on this together Update?

Renan reports: "I had a similar experience. HDFS works better with bigger files rather than many small files. What I did was to create a Map-Reduce job to reduce all those thousands of small files into only 17 big files, each of them containing all data for a given year [1998-2014].I didn't use Hadoop MapReduce for this, though. I used a different dataflow distributed engine that also implements map and reduce operators. I am working on providing Thiago these 17 big files. Once he gets the data, he can share them with you and explain how the data on each file are stored".

UNIMAS

Johari has a Raspberry Pi 2. It has double the RAM and a better processor. It will go to the Data Center in June.

Johari still has to uncover the problem of the traceroute from UNIMAS. UDP has been unblocked. The MYREN  host works fine and share most of the hops. Thus the problem must be in the first few hops.

UTM

Johari will contact Hanan to request someone to support PingER at UTM, now Saqib has left. (6/3/2015/. Hanan not replying to Johari, Johari will try another route to get a replacement for Saqib at UTM 8/12/2015).

The current case study is available in Google drive as a "Shared-PingER" document for review at https://drive.google.com/folderview?id=0B-NEKleLll79ZFNmUnhiVGJ0Nmc&usp=sharing_eid 

Saqib points out a number of Malaysian routes are IPv6 which could have problems for traceroute. Saqib is checking

MYREN

No update 8/12/2015

pingersonar-um.myren.net.my is down since may 31st.

Fizi had some problems with tracepath in https://perfsonar.myren.net.my/toolkit/gui/reverse_traceroute.cgi?function=tracepath. He has fixed it by changing the initial MTU.

NUST

Working with Dr Zaidi to get support. Last update last weekend. Await update from Dr Zaidi (8/12/2015)

PingER at SLAC

More improvements to  frequency.pl for full data (all possible pings, e.g. all 10 of the 10 pings each 30 mins) to speed up by a factor of 10, now can analyze all the raw data of 2015 between 2 hosts in < 30 secs. Try http://www-wanmon.slac.stanford.edu/cgi-wrap/frequency.pl?monitor=pinger.slac.stanford.edu&sites=pinger.unimas.my&begin_day=2&begin_month=1&begin_year=2015&end_day=6&end_month=8&end_year=2015&plot=freq&full=true&bin_width=0.01&metric=ipd&data=false

The Raspberry Pi at SLAC is still running smoothly since June 11th. It has been moved to the data-center so it now is attached to the same switch as pinger.slac.stanford.edu (to simplify comparisons). I am looking at monitoring its loopback NIC to compare with pinger.slac.stanford.edu monitoring its loopback to see the measuring host's impact on the RTTs.

We have a 4 host cloud with~16Gbytes each, plus access to 220Gbytes each. It is up and working. It was used for the paper submitted to NETAPPS1015. 

Working on the following hosts to be able to gather data 

Bebo arranged a meeting with the Colombia RENATA NREN folks and the minister of IT to discuss the use of PingER in Colombia. There is a web page at: Colombia. Les has sent an email asking them to install pinger2.pl at at least one site in Columbia. Sent a reminder email 2/27/2015. Bebo will send a gentle reminder to the RENATA people of Columbia to see whether they continue to be interested and need a meeting. They still seem interested.

Next meeting

Next meeting:  Wednesday Sept 2nd 2015 9:00pm Pacific Standard Time, Thursday Sep 3rd 2015  9:00am Pakistan time, Thursday Sep 3 2015 noon Malaysian time, Thursday Sep 3  2015 02:00am Rio Standard Time.  

Old Items

NUST/SEECS Pakistani PingER nodes status

Pink Background indicates host was bad last month, strike through says it is fixed, yellow is an new bad host.

Is it time to start paring down the list of PingER monitor hosts in Pakistan, starting with those that have been down for a while and despite your efforts they are not cooperating.  One might also look at the coverage by region in Pakistan and try and keep good coverage for all regions.

Traceroute at UTM 5/9/2015

The traceroute problem regarding maximum reachable hops ( i.e. 11 hopes ) may be since the Unix/Linux/OSX  traceroute uses UDP to send the requests. The first request is sent to a particular port (33434), with a ttl  to tell it how many hops to go to.  The ttl starts at 1 is incremented as it tries the next hop, also the port is incremented (up to 33465).  It looks like the first few UDP ports are enabled and then they are blocked. The Windows traceroute uses ICMP to send the probes so does not see the problem.

Raspberry Pi 5/9/2015

The two major issues with the Raspberry Pi would be:

• are the results statistically the same as for the other monitor at UNIMAS (e.g. use the Kolmogorov-Smirnov test); There is Advanced Project (Master by coursework student) working on the statistics of the data from the raspberry Pi and the production PingER monitor at UNIMAS to see how much they differ.
• is it reliable/robust is it clear what to do to debug problems remotely (e.g. if it is at Bario).  Looking at the monitoring data I have been unable to collect any from it (it is pingable, and port 80 responds, however the remote traceroute and ping_data.pl are not working) since Oct 20th which does not sound promising. Will need to evaluate the robustness of the unit by doing simulated scenario of various events such as power failure, hard and cold reboot, etc. Johari will need access to computer center to verify it comes up correctly after reboot etc.

If/when it works it would be instructive to look at the data from pinger and raspberry pi to Malaysia since the distances are shorter and the differences may show up better. For Sep-Oct 2014 when there was data measured from both Oct-Nov the averages for 20 paths was 52+-21ms (from pinger.unimas.my to 20 other Malaysian hosts) and 56+-21ms for raspberry pi to 20 other Malaysian hosts.

Linked Open Data

Cristiane reports (7/1/2015): "I am trying to automatize the triplification of PingER data on Kettle. For now, part of the transformation is made on Kettle and another is made by a Java code. Although this solution works for a data sample, is important to have the entire process on Kettle because it facilitates to understand, modify and control the triplification process."

Feb 2015

The plan is still the one seen before (see project proposal), experimenting those alternatives. Right now, they managed to triplify the data according to a new ontology that takes advantage of a combination of a current standard for multidimensional data (called data cube vocabulary) and a revised version of Renan's Moment ontology adaptation. With this we expect to have a better data organization than the previous solution.

They are now preparing a test plan (like a small benchmark) to be used on all alternatives so that we can compare the results accordingly. 

Aug 2014

Renan  finished the new pingerlod web site. The new thing is that it should be much easier now to modify the info texts. What Renan did was to put the texts into a separate file. The new version has been loaded on the server and some text added to describe how to use the map. However there is a bug that prevents it from executing the map. Renan reports that the bugs should be easy to fix. He has talked to his professor who suggested trying RDF Owlink, it should have faster responses to queries. Renan will research this.  It will probably mean reloading the PingER data so is a lot of work, hopefully this will improve performance. Before the rebuild he will make the fixes and provide a new WAR for us to load on pingerlod.slac.stanford.edu. He is also working on documentation (he has finished the ontology and has a nice interactive tool for visualizing it, since the ontology is the core of the data model of our semantic solution, this will be very helpful for anyone who uses our system, both a developer of the system and a possible user) and his thesis. Bebo pointed out that to get publicity and for people to know about the data, we will need to add pingerlod to lod.org.

Things he will soon do regarding documentation:

1. A task/process flow writing all java classes involved on all those batch jobs;
2. A Javadoc <http://www.oracle.com/technetwork/java/javase/documentation/index-jsp-135444.html> which will explain all classes and how they are used.

For the Linked Open Data / RDF which is in pre-alpha days, you can go to http://pingerlod.slac.stanford.edu. As can be seen this page is not ready for prime time. However the demos work as long as one carefully elects what to look at:

• Click on Visualizations, there are two choices:
  • Multiple Network Metrics: Click on the image: gives a form, choose from Node pinger.slac.stanford.edu pinging to www.ihep.ac.cn, time parameters yearly, 2006 2012, metrics throughput, Average RTT Packet loss and display format Plot graph, then click on submit. In a few seconds time series graph should come up. Mouse over to see details of values at each x value (year).
  • A mashup of network metrics x university metrics Click on image: gives another form, pinging from pinger.slac.stanford.edu, School metric number of students, time metric years 2006 2012, display format plot graph, click on submit. Longer wait, after about 35 seconds a google map should show up. Click on "Click for help." Area of dots = number of students, darkness of dots = throughput (lighter is better), inscribing circle color gives university type (public, private etc.) Click on circle for information on university etc.

• Renan will be working on providing documentation on the programs, in particular the install guide for the repository and web site etc. This will assist the person who takes this over. 

Renan is using OWLIM as RDF Repository. He is using an evaluation version right now. Renan looked into the price for OWLIM (that excellent RDF Database Management System he told us about). It would cost 1200EUR minimum  (~ 1620 USD, according to Google's rate for today) for a one time eternal license. It seems too expensive. No wonder it is so good. Anyhow, he heard about a different free alternative. Just not sure how good it would be for our PingER data. He will try it out and evaluate. He will also get a new evaluation of the free OWLIM lite.  

He has also made some modifications on the ontology of the project (under supervision of his professor in Rio) hence he  will have to modify the code to load the data accordingly.

Maria and Renan are advancing in some approaches to deal with PingER data, making it easier to be analyzed and integrated. In particular they have been busy studying and evaluating alternatives, analyzing results from the latest benchmarks on NoSQL (including RDF and graph based storages) database management, distributed processing and mediated  solutions over relational databases, and also other experiments with multidimensional analyses on Linked Data.  The new students involved are now understanding better the scenario and they have been interacting with Renan regularly. 

Cristiane has studied the PinGER data and how to cast it into Linked Open Data form. The size of the PingER hourly data for 1998-Sep 2014 archived via FTP in text form amounts to ~ 5.12GB and this corresponds to 15.66*10^9 (billion) triples. Then using 5  triples for each measurement and using Turtle without compression gives us 685 Gbytes or an inflation factor of ~ 200. 

When Christiane made the estimation of PingER triples, she wrote two documents that explain the process but they were in Portuguese. She has written the new versions in English.

• Counting PingER Measurements: https://www.dropbox.com/s/35itp7v6yasy3rb/Counting%20PingER%20Measurements%20_EnglishVersion.docx?dl=0
• PingER LOD Triples: https://www.dropbox.com/s/4oj5jqupwbujja5/PingERLOD%20Triples%20_EnglishVersion.docx?dl=0

Christiane's report is at: Size Inflation of PingER Data for use in PingER LOD

UM

Moved here 3/4/2015:

Ibrahim has setup distributed hadoop clusters. He has 2TB of disk space. Les has provided information on getting a subset of PingER data by anonymous ftp via ftp://ftp.slac.stanford.edu/users/cottrell.  It was put there last September. Information on how the data was put together is at https://confluence.slac.stanford.edu/display/IEPM/Archiving+PingER+data+by+tar+for+retrieval+by+anonymous+ftp. There is information on formatting etc at http://www-iepm.slac.stanford.edu/pinger/tools/retrievedata.html and some on the dataflows at https://confluence.slac.stanford.edu/display/IEPM/PingER+data+flow+at+SLAC. Renan at UFRJ has successfully used this data, he has also characterized the data in terms of bytes/metric per year etc.

Ibrahim has started downloading all zip files in the local machines. 6 weeks ago he downloaded 2 GB of Weather data to test his nodes cluster, he  wrote a simple Java program (Map, Reduce) to find the Average and it was working fine. 

Anjum reported that UM had experienced a TCP syn DOS attack prior to Mar 12th (when an IDS was put in place). It occurred mainly for several days before between the hours on noon- 2pm and 7-7 in the evening (Malaysia time). He suggested looking to see if PingER could spit the effect.  Ibrahim, Les and Anjum will look at. Les analyzed the data and sent it to Anjum

NUST

The following is from Samad 2/24/2015.

• buitms.seecs.edu.pk #We have to disable gathering data from this host because the person still don't want to continue with us as i have tried once again to convince him but the answer is same. Les has disabled from SLAC. 
• nukhimain.seecs.edu.pk # We were unable to gather data since 20th November, 2014 and now the Node is working fine and collecting data as well. 
• pinger.uettaxila.edu.pk #The node is working fine from last two weeks. 
• sau.seecs.edu.pk. #This Node is working fine now. 
• pingerjms.pern.edu.pk #This node is working now. 
• pinger.uet.edu.pk # this was also not working from so many days. and now its working fine and collecting data as well. 
• pinger.isra.edu.pk # This node is also working fine now.
• pingerlhr-pu.pern.edu.pk # This is also working fine now.
• pinger.kohat.edu.pk # Collecting data now.

• The IP of "pingerqta.pern.edu.pk" has been changed, Les has updated the databas at SLAC with the following

  DNS: pingerqta.pern.edu.pk 

  Old IP: 121.52.157.157 

  New IP: 121.52.157.148

Follow up from workshop

• Hossein Javedani of UTM is interested in anomalous event detection with PingER data. Information on this is available at https://confluence.slac.stanford.edu/display/IEPM/Event+Detection. We have sent him a couple of papers and how to access the PingER data. Hossein and Badrul have been put in contact. Is there an update Badrul?

The Next step in funding is to go for bigger research funding, such as LRGS or eScience. Such proposals must lead to publications in high quality journals. They will need an infrastructure such as the one we are building. We can use the upcoming workshop (1 specific session) to brainstorm and come up with such proposal. We need to do some groundwork before that as well. Johari will take the lead in putting together 1/2 page descriptions of the potential research projects. 

1. Need to identify a few key areas of research related to PingER Malaysia Initiative and this can be shared/publicized through the website. These might include using the infrastructure and data for: anomaly detection; correlation of performance across multiple routes; and for GeoLocation. Future projects as Les listed in Confluence herehttps://confluence.slac.stanford.edu/display/IEPM/Future+Projects can also be a good start and also Bebo's suggestion. 
2. Need to synchronize and share research proposals so as not to duplicate research works. how to share? Maybe not through the website, or maybe can create a member only section of the website to share sensitive data such as research proposal?

Anjum suggested Saqib,  Badrul and Johari put together a paper on user experiences with using the Internet in Malaysia as seen from Malaysian universities. In particular round trip time, losses, jitter, reliability, routing/peering, in particular anomalies, and the impact on VoIP, throughput etc.  It would be good to engage someone from MYREN.

Ibrahim

Ibrahim Abaker  is planning to work on a topic initially entitled " leveraging pingER big data with a modified pingtable for event-correlation and clustering".  Ibrahim has a proposal, see https://confluence.slac.stanford.edu/download/attachments/17162/leveraging+pingER+big+data+with+a+modified+pingtable+for+event-correlation+and+clustering.docx. Ibrahim reports 7/15/2014 "I have spent the last few months trying to understand the concept of big data storage and its retrieval as well as the traditional approach of storing RDF data. I have integrated a single hadoop cluster in our cloud. but for this project we need multiple clusters, which I have already discussed with Dr. Badrul and he will provide me with big storage for the experiment." No Update 8/20/2014.

"I have come up with initial proposed solution model. This model consists of several parts. The upper parts of the Figure below shows the data source, in which PingER data will be convert into RDF format. Then the data pre-processor will take care of converting RDF/XML into N-triples serialization formats using N-triples convertor module. This N-triple file of an RDF graph will be as an input and stores the triples in storage as a key value pair using MapReduce jobs"

Potential projects

See list of Projects