Adding support for L4S testing with TCP flows.

- Added support for Size based TCP sessions - Added support to configure a user-given device name to uniquely identify the device under test See the full changelog at https://excentis.odoo.com/knowledge/article/311

New: When the ByteBlower Endpoint application is allowed to communicate with the ByteBlower MeetingPoint while performing a test, the intermediate results are streamed to the ByteBlower MeetingPoint. This allows users (the API or the ByteBlower GUI) to monitor the results of the test while the test is still running. Please note, all ByteBlower components must be updated to at least 2.21.0 to enable this feature.

Improving UDP Stream and Trigger benchmark performance.

Feature: The ByteBlower Endpoint is now able to continue communicating with the MeetingPoint whilst running a test. This allows the user to stop a running test from the ByteBlower API or even from the ByteBlower GUI! GUI: The interface selection is now available on the top right of the first window. One will certainly recognise the gear icon!

- Allowing the user to select the traffic interface to use. This can be configured in the settings window accessable from the registration window. - UI: Moved the MeetingPoint address up, so it doesn't get hidden by the on screen keyboard.

Fixes: - In some cases the UI did not forward the MeetingPoint address to the backend application. This caused the application to stay in contacting. - The application could become unresponsive when the device went into deep sleep. This was caused by the application trying to catch up with the missed heartbeats.

- Fixed the behavior for flows with a source or destination port which is already in use. The WEP will now ignore the flow and report 100% loss. - Fixed a resource leak on Windows when running multiple consecutive scenarios. - Fixed a parsing issue detecting zero values, leading to bad parsing of MAC addresses, BSSIDs, .... - Enhanced the hearbeat processing thus avoiding ResponseTimeout issues.

Fixes - Heartbeats would fail if whitespace was given in front of at the end of the MeetingPoint address - The heartbeat protocol sometimes timed out before the full heartbeat was handled. Now the timeout is not triggered when the data is still going over the network. This fixes an issue where no results were available on slow and high-latency links. - The heartbeat protocol will now be less blocking when no network connection is available. - TCP Restart could reach a flapping state in slow/high-latency links, so the TCP Flow was restarted even before the TCP SYN/SYN-ACK/ACK mechanism finished.

Fixes: - Improved start time accuracy - Running TCP traffic in the downlink direction could stall after about 5 seconds - Running TCP traffic in the uplink direction could be very bursty - NetworkInfoMonitor snapshot interval is now correct. Features - The GUI now displays the device UUID and the MeetingPoint version when connected. - Added LatencyDistribution triggers

Our app did not render properly on iOS 14.4, this is fixed now.

Fix an issue where a stream used the UDP destination port as the UDP source port.

- We've revised the FrameBlasting Transmit approach. The new algorithm is better to match the configured duration and rate, especially in low bandwidth Wi-Fi environments - Fixed an issue where the device was marked as "Booting Device" when the device type contained a quote.

•Added support to collect the WiFi interface parameters over time: SSID/BSSID/Channel/RSSI/TxRate on supported devices.

Bugs fixed: * The Wireless Endpoint logged a lot of QML warnings in the console. * The Wireless Endpoint could report more frames to be transmitted than the actual number of frames transmitted in frameblasting flows. This can be detected by having loss in upstream UDP flows. * The Wireless Endpoint could report more frames being received than the actual number of frames received in frameblasting flows. This could be detected by having negative loss in downstream UDP flows.

- Fixing performance issue where TCP did not perform well on some devices - Fixing issue where TCP Upstream did not restart well

In the summer of 2018, we’ve introduced the option to automatically restart a broken TCP session. This feature provided continuous TCP traffic, even when the network itself encountered hiccups. At that time, this feature was only available on the ByteBlower server ports. In this first major release of 2019, we are bringing this feature to ByteBlower Wireless Endpoints. A big aid to test mobile devices even in difficult Wi-Fi environments.

- Fix issue with huge result-sets being communicated to the meetingpoint - Threads where created on the windows version and never destructed. This has been resolved. - Abort button didn't work anymore under certain conditions. This is resolved and you now can abort again the app

- Improvement on the core implementation. Prepared the code for TCP restart. - Improved heartbeat protocol. Where in previous versions, it could be so that your app stayed in contacting while there was WiFi or visa versa. This is now fixed and the app will handle much better the change between connected and disconnected

- build fixed for iOS

- Adding a command line interface - Minor bug fixes

- On some devices the app hangs when wifi is lost and doesn't recover. This has been fixed - Symantec Endpoint protection blocked our TCP traffic. We've altered the content on the TCP up so that Symantec Endpoint protection doesn't block it anymore

- Fix IPv6 UDP flows. The total transmitted/received bytes was incorrect. The calculation of the IP-header was wrong - Improved behaviour with channel hopping - Improved re-connection. Sometime WE didn't reconnect when WiFi signal was lost.

General improvements

La stabilité est améliorée pour les tests de longue durée