Knowledge base : Download Area > ByteBlower > Previous Releases > ByteBlower GUI

Version 2.9 of our popular ByteBlower® tool is ready to make IP testing faster and simpler than it already was! We’re on a constant quest to make ByteBlower more intuitive to deploy. So, we’ve added a range of new, practical features to the latest ByteBlower version. Your benefits? Faster, simpler and more efficient testing.

This version isn't available for download anymore. New versions can be found at

Changelog 2.9.8 (14th of Februari 2018)

  • RFC-2544 flows need to run for a minimum amount of time for decent results. Scenarios with less than the minimum will be marked faulty. In the solution view you'll find the suggested minimum time.

  • RFC-2544 flows in the report are now either black (good) or red (no results). The colour purple (result resolution) has been dropped, this warning wasn't very useful and couldn't be configured either way.

  • The total duration of RFC-2544 flows can be changed in the scenario view.

  • The FrameBlasting settings dialog has been made a bit smarter. It warns you of wrong burst configurations. Such an invalid config would make it impossible to edit the FrameBlasting template further.

  • ByteBlower ports can be docked again in the wizards. Whole tests can be created the wizards.

  • Fixed various edge cases while processing TCP flows. Especially the first and last results weren't always retrieved from the ByteBlower Server.

  • Corrected comments and headers in the CSV. The largest changes were in the TCP table and IPV4 ports sections.

  • Speed up generating reports with FrameBlasting flows on Wireless Endpoints. The MeetingPoint doesn't clear results, this caused for a lot of duplicate results requiring processing.

  • Configuring multiple TCP Flow templates with the same server port can trigger the PortAlreadyUsed exception. Due to an API change, this exception wasn't handled at the correct call.

  • To make it easier to debug test-runs, the message for a failed DHCP attempt now includes the ByteBlower interface where it goes wrong.

  • The decision whether a Wireless Endpoint has finished with its test is now delegated to the MeetingPoint. This helps especially when the test had no traffic or when it was aborted.

  • Added a fix for the case when only invalid latency packets are received.

Changelog 2.9.6 (14th of November 2018)

  • Fixed a nullpointer exception in the exception handling code. This error was triggered when sending multiple flows with undersized IPV6 frames.
  • Framesize modifiers now also work on ByteBlower Ports with an MTU smaller than 1500.

Changelog 2.9.4 (16th of October 2018)

  • In the CSV, jitter and invalid packetcount were added to the latency results over time. The documentation embedded in the CSV was spell-checked.

  • When the internal Eclipse browser isn't available, an external one is used. This helps Linux users who haven't yet installed webkitgtk-1.0-0.
  • The PCAP frame import wizard loads again. The code for this class wasn't loaded during startup.

  • Restarted TCP flows didn't stop after the right time. The TCP start/stop/reload code was simplified and debugged.

  • Corrected backwards compatibility for RFC-2544 flows. Projects created before 2.9.0 with now load with the corrected duration for this type of flows.

  • The CLT runs the complete batch again. A concurrency issue aborted the second scenario in the test.

Changelog 2.9.2 (19th of July 2018)

  • A fix for TCP flows with delayed start. The initial time to wait was not taken into account when aborting the flow.

Changelog 2.9.0 (3th of July 2018)

  • TCP test traffic is now restarted if it stops before its configured end. Reasons for such a restart are aborted TCP connections or no new received data. The reason is included in the warning section of the report. This behaviour is now enabled by default in new projects. But you can change this in the preferences.

  • You can control the way tables are sorted in the GUI. This is extremely useful in the Scenario view. By clicking on the column header, you can enable, disable or change the sorting order. The column used is shown in grey.

  • The scenario and batch view now include a duration column.

  • The status-line at the bottom of the GUI counts the selected items. This makes it easier to manage large ByteBlower projects.

  • Processing FrameBlasting results has been reworked. This new approach is faster resulting in a lower load on the client PC.

  • Starting the ByteBlower server from the right-click menu in the Server View includes its status now.

  • Wireless Endpoints on slow smart phones have a limited transmit timing accuracy. The report will show no warning anymore when less than one percent of all packets weren't transmitted.

  • The results over time require regular updates from the ByteBlower server. The report will notify you when not all intermediate results were processed in a large or long test. If the test is aborted early due to network connectivity issues, the report includes essential troubleshooting info.

  • The report includes a warning when a ByteBlower interface is asked to transmit more than physically feasible. In this statement you'll find which interface was overloaded.

  • The menu bar had an empty tab for custom wizards. This one has been removed, but ideas for new wizards are of course still welcome.

  • We currently don't support flows between two Wireless Endpoints. Such flows are now marked faulty and shown in the Solution View.

  • In the RFC-2544 wizard, a duration hint and additional info make the configuration much easier.

  • The RFC 2544 flow is editable now. Right-click and pick the jump to edit option.

  • Speedup of RFC-2544 benchmarks. They are now up to 2x faster than before. The main improvement was through better estimating of the final throughput.

  • For consistency, Auto HTTP method is renamed to Automatic HTTP method.

  • In the right click menu, Jump is renamed toJump to edit. The functionality remains the same, but we hope this name is more intuitive to new users.

  • We've updated the reporting engine (Jasper). This solves a number of rendering bugs in the library.

  • Fixed a couple rare null-pointer exceptions. One was due to a concurrency issue, the other was found in the UI logs.

  • Easier access to the system logs. You can open them with Window -> System info.

  • Projects created with 'Save As..' are now also found in the recent projects list in the File menu.

  • Some UI lag was noticeable when starting a test run. The root-cause was the Archive View. It is now faster and only updates the visible items.

  • Saved preferences weren't used in new projects. This bug has been fixed.

ByteBlower GUI version 2.5.6 brings you a number of new features and improvements. These make the user experience even smoother. Also, this version already contains some enhancements under the hood, to prepare for the upcoming release of the Wireless Endpoints. As an example, the docked information is now included in the report. This makes it even more straight-forward to rebuild test setups later on. In addition, TCP flows now offer the option to configure the TOS/DSCP byte. This offers additional support for testing policy-based routing. Below is an overview of all enhancements.


ByteBlower GUI 2.5 is no longer available for download. Please contact support if you thing you need this version

WARNING: A 2.5.x GUI only works with a 2.1.x server or higher. To be able to use all new features, we advice to update the server to the latest 2.5.x version. See the server download page for information how to install this server version. GUI 2.x can be installed side by side with an 1.x GUI.

WARNING: On Mac OS X, the 2.5.x GUI needs to be installed on version 10.10 (Yosemite) or later.

Changelog 2.5.6

  • Reporting
    • The CSV format has been updated. In addition to latency results over time, the reports also include frameblasting throughput values, a significant number of TCP statistics and the latency distribution results. Extra documentation is added in the CSV document to explain these new values.
  • User Interface
    • Improved robustness of the passport window when using the server support tool.
    • Code cleanup of the Realtime view. A number of exceptions are now resolved early, this results in a more readable logging file.
  • GUI Updater
    • The GUI updater adds a warning on failed updates. If the failure can't be corrected, a full reinstall is still necessary. In these cases, this recommendation is shown in the pop-up screen.
  • Test Runtime
    • Latency measurements are not yet supported on Wireless Endpoints. These measurements are also disabled for flows with a Wireless Endpoint configured as destination.
    • Fixed domain exception in the runtime. This error occurred while testing HTTP flows.
  • CLT
    • Fixed NumberFormatException in tests with Wireless Endpoints.

Changelog 2.5.4

  • User Interface
    • Resolving various edge cases when a MeetingPoint and ByteBlower server are accessed through the same address.
    • Improved the information mouse-over on wireless endpoints. The IPv4 address is added and the OS information is cleaner.
  • Reporting
    • To make it easier to distinguish the receive and congestion window, the colours were changed slightly. Regenerating old reports will also use these new colours.
    • The report includes the duration and throughput of FrameBlasting flows to Wireless Endpoints.
    • TCP flows failing to start and thus failing to negotiate the congestion window, won't show a congestion window.
    • The state of TCP flows failing to connect was not shown correctly.
    • The return path of the HTTP data flow consists mainly out of ACK-messages. In the Aggregate TCP results over time we hint to this direction with the name ACKS.
  • Test Runtime
    • Corrected the filter creating for flows from a Wireless Endpoint through a NAT.
    • NATDiscovery uses a probing packet to determine the NAT mappings. The name of the prototype frame is now included in this probing packet.
    • Verifying whether the Mobile TCP was able to connect to the server before requesting further statistics.
    • The Wireless Endpoints don't support Latency and Out Of Sequence measurements yet. This feature was disabled for them.

Changelog 2.5.2

  • General
    • Time calculations are based on GMT. This avoids problems in the underlying Java date libraries. This change fixes several issues for users in timezones at 30 minute offsets.

Changelog 2.5.0

  • User Interface
    • The TOS/DSCP byte is now configurable for TCP flows. This is feasible on ByteBlower servers running version 2.5.0 and up. For Frame Blasting flows, the TOS byte was already configurable. For user-friendliness, the TOS byte of all flows is listed together in the Flow View.
    • Servers with a recent update system (v3.3.0) can be updated through the GUI. This offers an additional possibility to update ByteBlower servers detached from the Internet.
    • To ease configuring long running tests, a duration can now also be specified in weeks.
  • Test Runtime
    • Reworked the filter generation algorithm. ByteBlower ports with different VLAN stacks and frames with non-IP frames are now even better supported.
    • Improved responsiveness of the runtime algorithm. Preparing support for the upcoming Wireless Endpoints.
  • Reporting
    • The docked info of the ByteBlower port is displayed in the ports table. It shows the interface the port was docked with additional information on the server itself. This feature makes it easier to recreate test setups.
    • The TOS byte of both TCP and FrameBlasting flows is included into the report.
  • Command Line Interface (CLI)
    • The printout is slightly altered to improve scripting tests. The CLI will display the name of the scenario and associated reports on standard output.
    • Batches configured in the project files can now be run with the CLI, using the '-batch' option.

We are happy to announce the availability of ByteBlower version 2.1.4 This brand new release enhances the new 2.0 experience with improved NAT reporting, additional TCP parameters and performance improvements. Here is an overview of what has changed:

Changelog 2.1.4

  • Runtime
    • Improved internal memory management while running a scenario.
  • UI
    • Fixed lay out bug in the Port View, that caused the MTU and Docked columns to be hidden.
    • The Realtime View now continues to update after closing and reopening it.
    • Hiding icons when not using DHCP or VLAN on a ByteBlower Port.
  • Reporting
    • By default, only HTML reports are generated now. PDF reports need to be enabled explicitly.

Changelog 2.1.2

  • Runtime
    • DHCP timeout overflow fixed
    • A NullpointerException was thrown when having 2 or more TCP Templates
    • In rare cases a exception was thrown during calculation of the Rx Filters
  • UI
    • Correct detection of 2.0.x servers. In previous version an c++ std::exception was thrown Now you will see incompatible version next to the server in the server-view
    • In Frame View, the layer 4 tab had no scroll bars
  • Reporting
    • Unit in the Aggregated TCP throughput was always kbps but the value was calculated according the report-setting. Now the unit will adapt accordingly the report-settings

Changelog 2.1.0


  • TCP enhancements
    • User may now edit the TCP rate limitation option that restricts the application layer data rate.
    • TCP slow start threshold (ssthresh) parameter is now configurable. The default value is 65535, and can be set as high as 2147483647 infinite. As a result, TCP flows can now reach their maximum throughput much faster in high-capacity networks. The value for existing projects is not changed. See this blog post for more information.
  • Aggregate results
    • Aggregate result tables return into reports. They contain aggregate TX and RX throughput information for each port in the scenario.
    • Aggregate TX (resp. RX) throughput is defined as the combined result of all flows sent (resp. received) on that port. It is calculated over both the alive time and the active time of a port.
    • The alive time of port is the period from the start of the first flow to the end of the last flow. The active time of a port is the sum of all periods with at least one active flow. Both periods are calculated per-flow and per-direction.
    • See Knowledge Base article for more information
  • Report enhancements
    • New NAT table showing the auto-discovered NAT port translations.


  • User interface enhancements
    • TCP rate limitation and slow start threshold are now configurable.
    • More intuitive editing of TCP duration field for both time and size based flows.
    • Added compatibility error when connecting to a 2.0.x ByteBlower server, since minor compatibility breaks were needed. If you see this message, you should update your server.
    • Flows and ports in the realtime view are now appropriately colored when in the error or warning state.
    • Columns are now automatically resized, to attain an optimal fit to display the content.
    • The Copy-Down algorithm is much faster now, which is noticeable when working with hundreds of objects.
    • Copy-Down incrementing is now possible on TCP port numbers and VLAN IDs too.
    • The first item of a table is now selected by default.
    • Adding scenarios to a batch sometimes failed 
  • User interface bugfixes
    • Fixed incorrect calculation of intended load when set in bytes per second.
    • Fixed minor bug in handling of time fields (e.g. start time of a flow within a scenario).
    • Ports created by wizards have now correct names.
    • The Throughput Wizard created one flow too much in some occasions.
    • The run button was visible even if no project was open.
  • Runtime bugfixes
    • TCP Congestion Avoidance Algorithm is now applied to the HTTP server as well as the HTTP client.
    • ARP was used for multicast and broadcast destinations.
    • IPv6 gateway address was sometimes ::
  • Reporting enhancements
    • Reports contain aggregate throughput results over all transmitted and received flows on each port.
    • Improved layout and subreport titles.
    • Chart legends are now on top of each chart, for an improved lay-out.
    • Time units are shown in the report
    • Negative loss is colored pink again


ByteBlower GUI 2.1.x is no longer available for download. Please contact our support desk of you think you need this specific version.

This release includes new features to the Wireless Endpoints. These support now Latency measurements (both directions) and IPv6 traffic. Next to this, finding past test runs has been made easier. We've added a search function to the Archive View and allow you to add tags to any test run.

Other improvements include a more detailed test report. It includes the TCP retransmissions over time. We've started adding suggestions based on measured results. In addition, the Archive and Server view have been made more user friendly.


This version isn't publicly available any-more. The latest GUI version is found in this article.

Contact If you do need this particular version.

Changelog 2.6.6 (2017-05-31)

  • Reporting

    • Reduced memory consumption. This makes multi-day tests with many Frame Blasting flows feasible.
      ByteBlower API
  • API

    • Scenarios with 100s of simultaneous flows had GUI crashes. This was caused by message corruption in the communication with the ByteBlower server. Proper synchronization fixed this issue.
    • The timeout when refreshing a MeetingPoint was longer than configured. This was caused by a conversion error. Now refreshing a ByteBlower Server or MeetingPoint takes at most 3 seconds.

Changelog 2.6.4 (2017-05-31)

  • Reporting

    • Fixed an error in drawing the Retransmission bar. This threw an exception when the Wireless Endpoint failed to prepare.
    • Report generation didn't take into account that a server might be unreachable. These servers are now listed with 'Unknown' as IPV4 address.
    • Corrected scaling of the retransmission bar in cancelled tests.

Changelog 2.6.2 (2017-04-21)

  • Wireless Endpoints

    • Increased delay for retrieving results on the Wireless Endpoint. This minimizes the chance on timeout exceptions when the device is not yet reachable.
    • Additionally clear the results over time after a test.
  • Reporting

    • Corrected the header of the TCP results over time in the CSV report.
    • The generated suggestions are added to the CSV report.
  • User Interface

    • Fixed bug in the Server View. The version info wasn't fetched correctly.
    • Tread-safe modifying the info in the Server View. This fixes a rare ConcurrentModificationException.

Changelog 2.6.0

  • Wireless Endpoints

    • With the new Wireless Endpoints (v1.1.0), it's now possible to include latency measurements. Both up- and downstream tests are supported. This can be configured in the flow view.
    • Also possible with new Wireless Endpoints is IPV6. Configuring such a test only requires docking an IPv6 port to a Wireless Endpoint.
  • Reporting

    • The info section of the report includes suggestions based on the test results. This makes it easier to debug your network.
    • The receive and congestion window have been merged into a single transmit window. This is the minimum of both, and is the value used by the TCP algorithm. This value is shown in a logarithmic scale.
    • The TCP retransmissions are shown at the bottom of the TCP results over time graph. As users note, drops of the transmit window and retransmissions often occur together.
    • The info section of the report includes the title of the test run and the used ByteBlower Servers and Meeting Points. This last detail includes their version and address.
    • The latency distribution additionally reports the relative amount of packets outside the measurement range.
  • Archive view and Run Title

    • The Run Title of each test is shown in the archive view. Right clicking a test allows one to change this title. Right clicking on a test-run allows one to change this title.
    • All reports of the same test run are stored in a folder with the name of the Run Title. This makes is easy to find all related tests. The reports are moved when changing the Run Title of a test.
    • The archive view includes a search field to quickly find previous tests results.
    • A number of tags are automatically inserted into each test run. As an example, there are tags inserted on tests with wireless endpoints, TCP and even latency measurements, or whether there was a significant amount of loss in the test. These tags give a quick overview of the test. They are displayed in the status bar at the bottom of the GUI. From the search field one can search on these tags. This makes it a piece of cake to find back your last latency test with a Wireless Endpoint.
    • Double-clicking a test opens the HTML report. This is a shortcut to immediately see its results.
    • The right-click menu in the Archive View allows one to delete tests or reports. This works both for single items and selections.
  • User interface

    • Reduced timeout when contacting a ByteBlower Server or Meeting Point. A refresh operation can also be canceled.
    • The available network interfaces are shown in the tooltip of a Wireless Endpoint.
    • A more fine-grained capability system has been added. This will block unsupported features, such as Out-Of-Sequence tests on an app.
    • Changed the interface to the result databases, It is now possible to have multiple GUIs open simultaneously.
    • The servers in the Server View are sorted on their configured address. Meeting Point and ByteBlower on the same server are now found together.
    • The TOS byte in the Frame View can be configured more easily through a number of preconfigured values. We have configurations for TOS, DiffServ(DSCP) and Wireless Media Extensions (WME)
    • The colouring of the Realtime View has been fine-tuned. This improves visibility on the throughput changes.
    • The ports section of the Realtime View includes more detail, especially for the Wireless Endpoints.
    • The right-click menu is a new, fast way to start a scenario or batch.
    • This welcome page includes a link to setting up your ByteBlower server (First steps, top)
  • Test Runtime

    • Recent apps (v1.1.0) support latency measurements. Outgoing tests support latency over time and latency distribution. A flow with an app as destination only allows latency over time.
    • Bugfixes to allow multiple ByteBlowers and MeetingPoints in a single test.



GUI version 2.4 features a wide range of improvements. A lot has been improved in the reports and the user experience has become even smoother. TCP charts and also the Latency charts in the reports contain much more valuable information now. The CSV and XLS reports contain detailed new Latency Results over Time tables. Also, throughout the entire GUI, the amount of required mouse clicks has been reduced, for increased ease of use. Below is a detailed overview of all enhancements.


ByteBlower GUI 2.4 is no longer available for download. Please contact our support desk if you think you need this specific version.

Changelog 2.4.4

  • User Interface
    • In the Frame View, new UDP and TCP Frames now have src port and dst port 4096 instead of 1024.
    • In the Frame View, the maximum allowed Frame Size is now 9000 bytes.
    • In the Port Forwarding View, the default Port Mapping is now 4096 to 4096 instead of 1024 to 1024.
    • Flexible IP address editing in the Port View. While editing, impossible values are allowed, but displayed in red.
  • Reporting
    • Mentioning Bitrate instead of Throughput.

Changelog 2.4.0

  • User Interface
    • Added shortcuts to create new items. For example, from within the Scenario View, you can now create new Flows when needed. This functionality has been introduced throughout the entire GUI.
    • When creating new Scenarios, one action is automatically added as well, reducing the amount of required mouse clicks. Similar features have been introduced throughout the entire GUI.
    • A default project name and author is now suggested when creating a new project.
    • In the Port View and in the Server View, the exact location where a ByteBlower Port has been docked is now displayed.
    • In the Server View, you can now start ByteBlower servers remotely, and also run the support tool.
    • New notification when an update is available for a ByteBlower Server.
  • Reporting
    • The charts with TCP results over time contain much more information now: Throughput, Goodput, Round Trip Time, Congestion Window and Receive Window.
    • The charts with aggregate TCP RX results over time contain much more detail now: The RX throughput contributed by each flow is now visualized for each port.
    • The CSV and XLS reports now contain new tables, displaying detailed Latency Results over Time.
    • On top of the average latency, the Latency charts now also show the minimum and the maximum latency over time, along with the jitter.
  • Runtime
    • When doing latency measurements, UDP checksums are now ignored. We introduced this as a workaround for a known bug in ByteBlower Server hardware.

GUI version 2.3 features a wide range of improvements. The most important ones are that you can now create VLAN Stacks (QinQ) and Frame Field Modifiers. The reports now feature new charts with Aggregate TCP Results. Also, the reporting engine has been improved a lot, so that reports of even huge, multi-day stability tests can now be generated very quickly. Below you will find an overview of all enhancements.

ByteBlower GUI 2.3 is no longer available for download. Please contact our support desk if you think you need this specific version.

WARNING: A 2.3.x GUI only works with a 2.1.x server or higher. To be able to use all new features, we advice to update the server to the latest 2.3.x version. See the server download page for information how to install this server version. GUI 2.x can be installed side by side with an 1.x GUI.

WARNING: On Mac OS X, the 2.3.x GUI needs to be installed on version 10.10 (Yosemite) or later.

Changelog 2.3.2

  • User Interface
    • The port view had a minor hickup when cutting a docked ByteBlower port. This has been resolved.
  • Runtime
    • Backwards compatibility with the past server releases is corrected. Especially on the 4100 ByteBlower models this can be noticeable.

Changelog 2.3.0

  • User Interface
    • VLAN Stacking. The VLAN View offers now the possibility to create VLAN Stacks (QinQ).
    • You can now enable a Frame Field Modifier on the payload of UDP Frames. Such a modifier can dynamically change a specific number of bytes in the UDP payload. Supported types : Incremental, Random and Unique.
    • In the Server View, interfaces now have the same name as the ones used in the API. (trunk-1-1, nontrunk-1, ...)
    • All default values are now also saved in the bbp project file. This will make it easier to maintain backward compatibility in future releases.
    • In the reports, the NAT table shows more details now about how the public ip address and port were discovered (ip => automatic discovery / manual, port => automatic discovery / port mapping)
    • Optimized updating of the Archive View.
    • Improved layout and background coloring of the Loss Preferences tab.
    • After a test is executed, the new item in the Archive View is now selected automatically.
  • Reporting
    • Featuring new charts, displaying the RX Aggregate TCP Throughput evolving over time for each ByteBlower port
    • Optimized reporting engine. Reports of even huge stability tests can now be generated very quickly.
    • Showing more info about active Port Forwarding.
    • HTML reports are now generated first. As soon as an HTML report is ready, it is displayed, while other reports like PDF, CSV, ... are being generated.
    • Showing all VLANs in each used stack. Marking the outer and inner ones.
    • When explicitly re-creating a report, but in the preferences no reports were marked, still create an HTML report.
    • Upgraded database library.
    • Preventing to generate multiple reports simultaneously.
    • Added warning when used ByteBlower servers need to be synchronized, to get aligned results over time.
  • Runtime
    • Improved memory management. Batches with hundreds of test Scenarios can now be executed without any problem.
    • Enhanced result database connection management.
    • The Realtime View is now cleaned up, and is hidden after a test finished.


ByteBlower GUI 2.2 is no longer available for download. Please contact our support desk of you think you need this specific version.

WARNING: A 2.2.x GUI only works with a 2.1.x server or higher. To be able to use all new features, we advice to update the server to the latest 2.2.x version! See the server download page for information how to install this server version. GUI 2.x can be installed side by side with an 1.x GUI.

WARNING: On Mac OS X, the 2.2.x GUI needs to be installed on version 10.10 (Yosemite) or later.

Changelog 2.2.10

  • User Interface
    • Fixed bug in the Frame View. Setting a manual IPv6 address is again possible.
    • Displaying user-friendly names in the Congestion Avoidance Algorithm column in the TCP View.
    • Fixed bug in the realtime results processor when no tcp traffic was received during an interval.

Changelog 2.2.8

  • Reporting
    • Optimized report generator. Reports are generated much faster now.
    • HTML reports are generated first, and are opened immediately while other formats are being generated.
  • User Interface
    • The Progress View shows now which type of report is being generated. (html/pdf/csv...)
    • The Archive View dynamically updates itself while new tests are executed or reports are created.

Changelog 2.2.6

  • User Interface
    • Closing the GUI while generating a report is not causing an error anymore.
    • Improved cleanup handling of canceled HTTP flows.
    • Importing Frames from pcap files is working again.
    • Allowing to send a mix of IPv4 and ARP packets.
    • Removed the possibility to override discovered NAT addresses in the Frame View on Layer 4, because this was confusing.

  • Runtime
    • Optimized data communication for realtime results.

Changelog 2.2.4

  • User Interface
    • In the Port View, NAT can now be turned off again.
    • When the Scenario Pause dialog appears, you can now still access the entire GUI.
    • Proper disposal of closed views. Input not set any more on disposed widgets.
    • Added missing menu Icons on Windows.
    • Executed Test meta data will immediately be persisted when a test is started. Now, when you are running a long stab test, and somewhere in the middle there is a power outage, you can still generate reports from the gathered test data up until then.
    • Removed unused Batch Initialization Time preference page.
    • Fixed estimated test end time in the Progress View.

  • Reporting
    • Assuring that the aggregate TCP throughput is using the last available data.
    • Improved results over time chart when the duration of snapshots is bigger than the resolution. This is only noticeable for very short tests, where all points are clearly distinguishable.
    • Fixed bug when generating charts containing empty series. For example when eavesdroppers received no traffic.
  • Runtime
    • Generating bursty traffic is possible again, because of a bugfix in the ByteBlower server software.

Changelog 2.2.2

  • User Interface
    • Port Forwarding now also works when NAT was first set to No.
    • ByteBlower Ports can be deleted again.
    • Fixed desktop icon on Windows.
    • Rounded view tabs.

  • Reporting
    • Fixed positioning of the first second of results over time charts in the report.
    • Latency distribution charts are no longer hidden.

Changelog 2.2.0

  • User interface enhancements
    • The New Port Forwarding View allows for fine-grained configuration of port mapping rules.
    • The Realtime View now features automatic highlighting of updated values.
    • The file menu includes links to recent projects.
    • When only a single scenario/batch is available, the quick run buttons will allow for these to be started immediately.
    • New  Cubic TCP Congestion Avoidance Algorithms.
    • A pause can be configured between initializing and running the test.
  • User interface bugfixes
    • Fixed Jump functionality.
  • Reporting
    • Optimized report generator. Huge reports can now be generated, without using lots of memory and cpu.
    • Aligned charts, so they can easily be compared.
    • Fixed page length for PDF reports.
  • Runtime enhancements
    • Improved handling of realtime test results, resulting in reduced memory footprint and enabling multi-day endurance tests.
    • Port forwarding functionality.
    • Reduced management network overhead by aggregating interaction with the ByteBlower server.
    • Asynchronous DHCP leads to much faster initialization times.
    • Faster and more robust NAT discovery.
    • Caching resolved MAC addresses.
    • A configurable grace period limits the time to wait for HTTP sessions to die.


We are pleased to announce the immediate release of the ByteBlower GUI 2.0.

This first installment of our 2.x GUI series brings highly anticipated features to ByteBlower and prepares our platform for the future. The most important features and functional changes from the 1.x series are listed below.

Please note that a 2.x GUI can only work with 2.x servers! Upgrading a server from 1.x to 2.x requires a reinstallation. This process is described on the 2.x server download page.

While the 2.0 GUI is not compatible with older 1.x servers, it is able to detect them. Furthermore, a 1.x and 2.x GUI can be installed side by side, so it is possible to work with all servers (old and new) from a single client device!


ByteBlower 2.0 brings a ton of changes to ByteBlower. The list below only contains the most important new features and the functional changes between the 1.x and 2.x GUI. Changelogs of the 2.0.x patch releases are listed at the bottom.


  • Test data gathering engine
    • Test results (such as periodic flow results) are now constantly gathered while a test is running.
    • Test data is stored in file-based databases (one per test). These databases are located in <home-directory>/byteblower/workspace_v2/TestData.
    • Test execution is thus decoupled from reporting, since the test data is persisted.
  • Project can be configured while test is executing
    • Before a test a deep copy of the complete project is taken.
    • User can work on opened project or open another project. The runtime component uses its own version.
  • Broadcast flow destination behaviour is changed
    • Broadcast destination now functions as a 'fixed' address, just like the unicast destination.
    • In addition, all ports (except the source port) in the project are implicitly added as flow eavesdroppers during scenario initialization.
  • Batches and batch items are now easier to use
    • Absolute start times are no longer supported. When opening an 1.x project file, all items with absolute timing are automatically converted to a relative time (i.e. after the previous scenario) of zero.
    • The initialization time property (which acted as a minimum initialization duration) of a batch item is removed.
  • Optimization of project model's internal refresh algorithm. GUI is more responsive and large-scale projects can be created more easily.
  • TCP default settings have changed
    • ServerDefault is removed as a TCP congestion avoidance algorithm. The new default value is SACK.
    • Window scaling is enabled by default with a default scaling value of 2.
  • Interoperability with 1.x servers: refreshed 1.x servers are shown as incompatible servers in the configuration view. They also show their (incompatible) 1.x version.
  • Server auto-detection service has been removed.

User interface

  • New Realtime view
    • Shows dynamic information about ports and flows, such as IP addresses obtained by DHCP.
    • Shows real-time flow results (throughput, frame rate, latency, ...)
  • Revamped archive view
    • Based on the test data gathering engine.
    • Allows new reports to be generated from old test data (e.g. after a GUI update with report bugfixes or enhancements).
  • Real-time port information is removed from the Port view; it is shown in the Realtime view instead.
  • The subpanels TCP and Frame Blasting of the Flow Template view have been separated into two views of their own.


  • New reporting engine
    • Reports are restyled but have a familiar structure.
    • Flexible reporting engine and test data stored in the database allows for future customization and enhancements.
  • Result over time charts
    • Show flow properties over its duration.
    • Charts available for throughput, latency and out of sequence.
  • Error and warning messages include a timestamp and are more verbose. For example, if some frames are indistinguishable or identical, these are listed in the warning message.

Known issues

  • General
    • Command-line tool is not yet available.
    • Latency on the 2100 still requires the user to set the UDP checksum to zero (see KB article).
    • Pause after DHCP feature is disabled. Scenario will always be started immediately after initialization.
    • Runtime port properties (such as IP addresses) are retrieved immediately after a port is configured. In the case of DHCPv6, DCHPv6 may be ready before a Router Advertisement was received. This currently results in a :: (i.e. zero) gateway address.
    • The TCP congestion avoidance algorithm is only set on the HTTP client. For flows using HTTP GET (the default case), this means that setting a congestion avoidance algorithm has no effect, since the HTTP server is the source.
  • User interface
    • File menu: No recently opened projects.
    • File menu: No importing from or exporting to ByteBlower frame (.bbf) files. Importing from PCAP (.pcap) files is still possible.
    • Wizards: Not all Wizard content may be visible on some (Gnome-based) Linux operating systems, such as Fedora.
    • Log view is removed from the GUI. It may return in another form.
  • Reporting
    • Result over Time graphs are not aligned.
    • Aggregate results are missing.

2.0.8 changelog

  • User interface bugfixes
    • A compatibility error is shown when connecting from 2.0.x GUI to a 2.1.x (or higher) ByteBlower server, since minor compatibility breaks were needed between the two. This is backported into the 2.0.x software to remind te user to update the GUI application.

2.0.6 changelog

  • Execution bugfixes
    • Fixed important TCP window scaling issue, which caused TCP flows with different window scaling settings to all take the same settings. [requires server 2.0.6]
    • TCP window settings are now configured on both HTTP client and server. This makes sure HTTP servers (re)used in multiple flows as both sender (GET) and receiver (PUT) always contain the correct settings.
    • Identical and indistinguishable frame detection is now restricted to a single physical interface, so warnings only show up when appropriate.
    • Fixed minor DHCP issue that could cause failed DHCP attempts to be interpreted as timeout issues instead of initialization errors. [requires server 2.0.6]
  • User interface bugfixes
    • New-Delete-Cut-Copy-Paste actions are now correctly enabled and disabled in the expected situations throughout the GUI.
    • Fixed issue where the global undo/redo operations were not available in the Realtime view.
    • Server view is now updated whenever a port is renamed or its status (i.e. ok or error) changes.
    • Central part of the Server view can no longer be resized.
    • Zero MAC address are no longer supported and are marked as invalid.
  • Reporting bugfixes
    • Reports now correctly show 'Error' (instead of 'Executing') as test status when something went wrong unexpectedly during execution.

2.0.4 changelog

  • Execution bugfixes
    • Creating ByteBlower ports on 2100 or 4100 servers with more than one trunking interface should no longer gives an opaque error with message 'Technical error: map::at'. [requires server 2.0.4]
    • Fixed unexpected Identical Frames notification when multiple source ports are behind a NAT device.
  • User interface bugfixes
    • Fixed NullPointer issue that sometimes occured after a UI component is disposed.
    • Wizards: New servers can now be created from within the Wizards again.
    • Scenario view: Time Last Frame Sent value now shows '-' instead of an invalid value when a flow's duration is less than the interframe gap (and 0 packets are sent).

2.0.2 changelog

  • Execution bugfixes
    • Fixed issue where ByteBlower ports were sometimes created on a wrong server, when multiple servers were used in a single scenario.
    • DHCP client now always asks for default gateway and netmask values explicitly, through DHCP option 55 (Parameter Request List). Because of this, it now works with DHCP servers that don't provide these values by default. [requires server 2.0.2]
    • Fixed issue where server could crash when a scenario is canceled. [requires server 2.0.2]
  • User interface bugfixes
    • Run scenario and Run batch buttons now become enabled (resp. disabled) when the selected scenario or batch becomes valid (resp. invalid).
    • Starting a batch opens the Realtime and Progress views, just like when starting a scenario.
    • When changing the frame repeat value in the Frame Blasting view, the frame and byte weights are now updated immediately.
    • Realtime view correctly shows pending instead of connection as status when a scheduled TCP flow has not yet started.
    • Progress view takes start times of size-based TCP flows into account as well, when calculating the scenario duration.
    • Various fixes to copy down and copy down incrementing behavior (IPv4 vs. IPv6, multicast addresses vs. multicast source addresses, ...).
    • Improved the UI refresh algorithm. Fixes issue where views were not updated instantly on Windows and generally improves the GUI performance.
  • Reporting bugfixes
    • Generating a report for a flow with a fixed destiation no longer fails.
    • Multiple latency distribution charts in a single report will no longer overlap.
    • Reports now show the project version instead of the client software version. This version is no longer (incorrectly) 1.99.0.
    • Fixed styling of some fields in the report.
    • Configuring an MTU above the physical maximum now results in a friendlier error.


ByteBlower GUI 2.0 is no longer available for download. Please contact our support desk of you think you need this specific version.


We to help you!