Knowledge base : Knowledge Base > ByteBlower
 

2.20.2   (2023-04-03)

Fixes

When the Scenario Heartbeat interval was set after Prepare() was executed on

an Endpoint, the new Scenario Heartbeat interval value was not propagated to the Endpoint.

The MeetingPoint could mark an Endpoint as "unavailable" when a test longer

than 1 minute was running.

This could lead to the ByteBlower GUI stopping a scenario while the test was

(rightfully) still running

2.20.0   (2023-02-27)

New features 

  • The server now provides latency histograms over time when using the latency distribution trigger. This allows to have latency histograms per configured measurement interval.   
    This feature is useful when measuring the QeD performance of a specific flow. 

Enhancements 

  • The TPID (EtherType) of a VLAN tag can now be configured through the API. 
  • Faster filtering for traffic containing multiple VLAN tags. 
  • The ByteBlower 2100 and ByteBlower 4100 server series did not support 4 stacked VLAN tags as the other series did.  Now all systems have the same limit. 
  • The MeetingPoint will now automatically query the ByteBlower Endpoint for the test results when the Endpoint returns after a test. 
  • The MeetingPoint needed up to 5 seconds for initial registration of an Endpoint.  The MeetingPoint now temporarily lowers the communication interval so this procedure finishes earlier. 
  • The MeetingPoint tagged unverified ByteBlower Endpoint platforms as “Unsupported”.  The tag is renamed to “Unverified”, which means Excentis did not have tested this platform. 

Bugs fixed 

  • The MeetingPoint stopped accepting API calls when an “Unavailable” ByteBlower Endpoint was addressed.  This did not only influence the calling test, but also effected other tests and users with “Available” Endpoints.  The only remedy for this was restarting the ByteBlower MeetingPoint service on the ByteBlower server. 

In 2020, the decision was taken by the Python community to stop support for Python 2.x versions.

However, in order to make life easier for some of our customers,  we have continued supporting Python 2.7 for API users who may have chosen not to upgrade their Python environments.

However, we will stop supporting Python 2.7 after the next API release.

We are discontinuing support because it has now been three years since Python ended support for this version which means that most users have moved away from Python 2.x.

This will better help us develop and improve the Python API.

If you have any questions about this, please contact us at support.byteblower@excentis.com

We would like to inform our customers that ByteBlower 2100 and 4100 servers are no longer available.

This was decided due to declining interest in these systems and the alternative systems (5000 & 3000 series) we provide.

Even though they are no longer available for purchase we understand that some of our customers still use these systems, and therefore we will continue support until the end of 2024.

This means all software features that are released in this period will apply to both systems as well as the other series in our range.

If you have any questions about this, please contact us at support.byteblower@excentis.com 

2.20.0  (2023-02-14)

Enhancements:

Increased robustness of RFC-2544 for WEP

  • RFC2544 is a widely known benchmark. The ByteBlower GUI implements the throughput measurement.
  • In the past, tests would more likely fail due to the stress the test places on the network.
  • RFC-2544 places a heavy load on the system. For fail-safe tests it’s recommended to add route the WEP mgmt. traffic over a different interface like e.g., the golden client.

UI improvements in the ByteBlower server updater. 

When errors happen resuting in failed tests, we have implemented improved logging which makes the feedback of the server more visible.

  • Improved logging of error-cases from the ByteBlower Server.
  • More efficient debugging as as a result.
  • Some examples of error messages
    • Code
    • Improper license
    • Broken updater that was fixed in later releases.

 

Improved default TCP setting.

  • New default value for TCP Slow Start Threshold: 2147483647 (=Integer.MAX_VALUE) (=infinite). The old default was 65535 (=0xFFFF).
  • OS behaviour is closer to infinite value.

CCDF Graph

An empty CCDF graph is now drawn in the ByteBlower report when all packets are below the latency measurement range. This can happen due to negative latencies of a Wireless Endpoint. See also our Golden client KB article to perform such tests. 

Changelog for 2.20.0 

 

New features 

  • Adding latency histograms over time.  This can be used to visualise/verify the QeD performance of a specific application. 

Enhancements 

  • Adding additional frameblasting interval statistics such as : 
  • When was the first frame sent or received 
  • When was time last frame sent or received 
  • Minimum and maximum size of the frames received 
  • Better latency measurements for cumulative snapshots.   
    The Endpoint now keeps track of the cumulative data itself instead of delegating this to the MeetingPoint. 

Fixes 

  • Fixing crashes when a management/traffic interface was selected, but no IP address was available. 

It has recently come to our attention that older firmware versions of the Netgear M4300 can potentially cause throughput issues with 2.5Gb/ 5Gb links.

Here we see a TCP flow with many retransmissions and a non-optimal throughput.

In order to resolve this issue, the switch should be upgraded to the latest firmware version.

1) Go to https://www.netgear.com/support/product/m4300-96x.aspx#download

  • Download firmware

2) Log in to your Netgear switch via the web CLI.

  • The default IP is 192.168.0.100
  • Username: admin
  • Password excentis

3) Upload firmware.

  • Go to 'Maintenance' → 'HTTP File Upgrade'