tcpexmachina / remy Goto Github PK

I don't know how to run it. When I followed the steps in README and run ./remy, I got like this:

An input configuration protobuf must be provided via the cf= option.
You can generate one using './configuration'._

please tell me what should I do to run remy and what is 'option' for.
I am really new to it. Sorry to interrupt. I will appreciate your reply, thank you!

I find that include guards like "PACKET_HH" and "UTILITY_HH" are too short for the safe reuse of your header files.

when i compile (make phrase) on this project, compile went wrong, and my operation is MacOS
git checkout v0.1

whiskertree.cc:188:10: error: use of undeclared identifier 'accumulate'
  return accumulate( _children.begin(),
         ^
1 error generated.
make[2]: *** [whiskertree.o] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

thanks

I try to compile remy with boost 1.55.it comes error

../boost/utility/enable_if.hpp:59:10: error: base class ‘struct boost::disable_if_c<true, void>’ has a non-virtual destructor [-Werror=effc++]

then i try boost 1.50.another error

sendergang.hh:53:9: error: ‘class SenderGang<Rat>::TimeSwitchedSender’ has virtual functions and accessible non-virtual destructor [-Werror=non-virtual-dtor]

Any suggestion?

I have some errors related to undefined references to functions. Following is the complete error log:

sender-logger.o: In function SimulationResults<WhiskerTree> run_simulation_for_results<Rat, WhiskerTree>(WhiskerTree&, NetConfig&, Rat&, unsigned int, unsigned int, double)': /home/collosalkid/Documents/tcp_exMachina_remy/remy/src/sender-logger.cc:36: undefined reference to SimulationResults::add_run_data(NetConfig const&, double)'
sender-logger.o: In function SimulationResults<FinTree> run_simulation_for_results<Fish, FinTree>(FinTree&, NetConfig&, Fish&, unsigned int, unsigned int, double)': /home/collosalkid/Documents/tcp_exMachina_remy/remy/src/sender-logger.cc:36: undefined reference to SimulationResults::add_run_data(NetConfig const&, double)'
sender-logger.o: In function void serialize_to_file<FinTree>(SimulationResults<FinTree>&, std::string&)': /home/collosalkid/Documents/tcp_exMachina_remy/remy/src/sender-logger.cc:60: undefined reference to SimulationResults::DNA() const'
sender-logger.o: In function void serialize_to_file<WhiskerTree>(SimulationResults<WhiskerTree>&, std::string&)': /home/collosalkid/Documents/tcp_exMachina_remy/remy/src/sender-logger.cc:60: undefined reference to SimulationResults::DNA() const'
collect2: error: ld returned 1 exit status

I checked the undefined error relating to "SimulationResults::add_run_data(NetConfig const&, double) " and found that there is no class SimulationResults as the class template of SimulationResults is defined to . I am probably missing something here so please let me know how to resolve this.

Note: I have removed the flags -Werror and -Weffc++ from all makefiles as it was giving some warnings as errors.
Thankyou

Currently there is no license file, or license attached to the source. As a result this makes this project cool, but can't look at it for fear of implementing something which is protected.

I was trying to run remy source code using intructions as provided in readme. I am using protobuf 3.5.1 version and Ubuntu version is 14.04. When I ran the 'make' command after running ./autogen.sh and ./configure repectively, i received this error:

In file included from configrange.hh:4:0,
from evaluator.cc:3:
../protobufs/dna.pb.h:4210:20: error: base class ‘struct google::protobuf::internal::integral_constant<bool, true>’ has a non-virtual destructor [-Werror=effc++]
template <> struct is_proto_enum< ::RemyBuffers::MemoryRange_Axis> : ::google::protobuf::internal::true_type {};

Please help me resolve this error.
Thankyou

the website speaks of hoping that Remy performs in real world cases as it does in the simulations... but I assume by now real world test cases have been done? Is anyone using Remy derived CC algorithms in production?

Would you like to add the configuration script "AX_PTHREAD" to your build specification?

I want to run a Remy-optimized TCP stack in a widel-deployed application. We could do this on top of UDP-- as most non-TCP transport protocols operate now.

• Do you foresee any problems with this?
• What would be the easiest route?
• Would anyone else be interested?

When using only one congestion signal, the following assertion error occurs after some time regardless of the network parameters (not exhaustively tested):

remy: memoryrange.cc:25: std::vector<MemoryRange> MemoryRange::bisect() const: Assertion '!(ersatz_lower == x._upper)' failed.

It seems to proceed normally without this assertion, although I am unsure of the implications.

1. Links should have a buffer with a default value of infinity. DONE
2. Packets must have a sequence number, which is an int, not uint. DONE
3. ACKs should be treated as cumulative, even without losses. DONE
4. RatBreeder::improve needs to be broken up into smaller fragments: promote_whisker, optimize_whisker. Also, remove needless promotion at the end. generation % 4 == 0 is misleading. DONE
5. Change interface to packets_received to support Aimd and Rat. DONE
6. Aimd should be on the main branch? DONE
7. slowewma must be optional. DONE
8. Rat should be loss resilient. DONE
9. Close other pull requests. DONE
10. Add jemalloc
11. SenderGangOfGangs refactoring, and create templates with SenderType1 and SenderType2 everywhere. DONE
12. Maybe mention congestion signals in the training/testing/provenance whatever.
13. Try and integrate ByteSwitched and TimeSwitched Sender into the main branch. DONE

As you told me to, I build remy inside travis CI. It gives me errors on the command
sudo apt-get install -y --allow-unauthenticated libglfw3-dev as written in .travis.yml file. The main error is E: Sub-process /usr/bin/dpkg returned an error code. Here is a link to the log.
https://api.travis-ci.org/v3/job/357396501/log.txt
I googled the error and tried sudo apt-get upgrade, sudo apt-get clean but didn't get any useful results.
Thanks