Inheritance diagram for ns3::Bug555TestCase:

Collaboration diagram for ns3::Bug555TestCase:

Public Member Functions
	Bug555TestCase ()
virtual void	DoRun (void)
	Implementation to actually run this TestCase.
Public Member Functions inherited from ns3::TestCase
virtual	~TestCase ()

Private Member Functions
void	NotifyPhyTxBegin (Ptr< const Packet > p)
void	SendOnePacket (Ptr< WifiNetDevice > dev)

Private Attributes
Time	m_firstTransmissionTime
ObjectFactory	m_mac
ObjectFactory	m_manager
unsigned int	m_numSentPackets
ObjectFactory	m_propDelay
Time	m_secondTransmissionTime

Additional Inherited Members
Public Types inherited from ns3::TestCase
enum	TestDuration { QUICK = 1, EXTENSIVE = 2, TAKES_FOREVER = 3 }
	How long the test takes to execute. More...
Protected Member Functions inherited from ns3::TestCase
	TestCase (std::string name)
void	AddTestCase (TestCase *testCase) NS_DEPRECATED
	Add an individual child TestCase case to this TestCase.
void	AddTestCase (TestCase *testCase, enum TestDuration duration)
	Add an individual child TestCase to this test suite.
std::string	CreateDataDirFilename (std::string filename)
std::string	CreateTempDirFilename (std::string filename)
bool	GetErrorStatus (void) const NS_DEPRECATED
std::string	GetName (void) const
bool	IsStatusFailure (void) const
bool	IsStatusSuccess (void) const
bool	MustAssertOnFailure (void) const
bool	MustContinueOnFailure (void) const
void	ReportTestFailure (std::string cond, std::string actual, std::string limit, std::string message, std::string file, int32_t line)
void	SetDataDir (std::string directory)

Detailed Description

Make sure that when multiple broadcast packets are queued on the same device in a short succession no virtual collision occurs

The observed behavior is that the first frame will be sent immediately, and the frames are spaced by (backoff + SIFS + AIFS) time intervals (where backoff is a random number of slot sizes up to maximum CW)

The following test case should not generate virtual collision for the second frame. The seed and run numbers were pick such that the second frame gets backoff = 0.

                 frame 1, frame 2
                 arrive                SIFS + AIFS
                 V                    <----------->

time |-----------—|----------------—|----------—|--------—> 0 1s 1.001408s 1.001442s ^ ^ ^ start TX finish TX start TX frame 1 frame 1 frame 2 ^ frame 2 backoff = 0

The buggy behavior was that frame 2 experiences a virtual collision and re-selects the backoff again. As a result, the actual backoff experience by frame 2 is less likely to be 0 since that would require two successions of 0 backoff (one that generates the virtual collision and one after the virtual collision).