Mark Dranse

RIPEstat Live Demo #4

Mark Dranse

5 min read

6 You have liked this article 0 times.
0

This article summarises yesterday's RIPEstat demo session, including new features, changes and improvements. At the end of the article you can find the date of the next demo, and options for providing feedback.


For reference, please see summaries of the previous demo sessions:

During this fourth demo the main improvements and new features were presented. More than 30 attendees joined the session.

Main Improvements

  • performance and stability has been improved significantly
    • improvements through more efficient code (connection pooling, 
smart data loading)
    • usage of faster backends
    • better, refactored libraries
    • analysis page for plugin performance (mainly for developer debugging)
    • module 'time to load' estimations refined
  • more informative error messages
  • module reload capability and intelligent timeouts
  • structure for direct links has been changed to be shorter, more meaningful and easier to share, e.g.
    • old URL: .../[{input.BlacklistInput:+{prefix:+193.0.0.0/24}}...
    • new URL: .../blacklist?value=193.0.0.0/24
  • methodology text has been improved throughout

Improvements on plugins

  • IP prefix table has been improved
    • the average number of peers is shown for each entry
    • time gap aggregation is used to reduce data noise (data is aggregated to a greater extent as it gets older)
    • a time period selection filter has been added
  •  whois plugin shows more helpful information
    • greater focus on most descriptive data
    • less informative data is hidden
    • queries for resources outside of the RIPE region redirect to the responsible RIR whois server 
  •  blacklist plugin has been improved
    • shows percentage of prefix blacklisted
    • it is clearer now for what time range we have data
  •  spider graph showing AS path length has been improved
    • opacity added to highlighted area
  • improved output of AS origin histories
    • isolated withdrawals are now ignored which removes empty lines showing up in the graph 
The concept of the future design was presented and can be seen under http://stat.ripe.net/media/new_design.png .

Questions and Answers

Q1: The data in the design concept screen shot seems to be incorrect?

A1: Yes, that can easily be. The screen shot comes directly from a graphic program and is based on dummy data.

Q2: What is the geolocation data based on? And is it possible for people to adjust their location on the map?

A2: The data is based on what Maxmind provides. At the moment it is not possible for us to adjust the information, we can only use it as is. However, we have started to reach out to geolocation providers and are currently discussing with them how they can make their data more accurate. As an interim solution, we could implement a comments field as part of the geolocation box to allow people to leave a note if their location is not accurate. It is planned to implement comment fields for all boxes, not just for the entire page.

Other comments:

  • Good that performance has been improved
  • Additional data filters are useful
  • The yellow on the second page is a bit difficult to read

Feedback

As always, we invite your comments, questions and feedback on our demo session and stat.ripe.net itself.
The main place for questions and discussions is the Measurement, Analysis and Tools Working Group:  mat-wg@ripe.net ."> mat-wg@ripe.net .

You can also use one of the following channels:

Video

The video archive of the demo is available from the RIPEstat website, with direct links here for your convenience:
 

Future Demo

Our next demo is scheduled for 17 May 2011. The link will be announced on RIPE Labs closer to the date.
 
You can also hear more about RIPEstat at the MAT-WG during RIPE 62 which will be held in Amsterdam from 2-6th May 2011.

 

 

 

 

 

6 You have liked this article 0 times.
0

About the author

Comments 0