First, use a distinct sequence numbers for each packet. When one of the receivers notices a missed packet, it asks sender to resend (to all receivers).
As an optimization, use bigger chunks. Use a window of packets. If the transmission is reliable, then expand the window size, so each sequence number covers to a (much) larger chunk of packets.
These are the basic reliability techniques of TCP. Reliable multicast borrows these from TCP. Note real TCP isn't usable for multicast as each TCP transmission has exactly one sender and one receiver. I think entire TCP protocol is based on that premise.
My main blog
Labels
_fuxi
(75)
_IV
(146)
_misc
(5)
{610610
(30)
algo
(1)
automatedTrading
(8)
banking/economy
(3)
book
(14)
c++misc
(125)
c++real
(15)
c++STL/java_container
(7)
cppTemplate
(1)
db
(13)
DB_tuning
(4)
deepUnder
(1)
dotnet
(69)
eTip
(17)
excelVBA
(12)
finance+sys
(34)
financeMisc
(24)
financeRisk
(2)
financeTechMisc
(4)
financeVol
(21)
finmath
(17)
fixedIncome
(25)
forex
(16)
IDE
(24)
invest
(1)
java
(43)
latency
(4)
LinearAlgebra
(3)
math
(30)
matlab
(24)
memoryMgmt
(11)
metaPrograming
(2)
MOM
(15)
msfm
(1)
murex
(4)
nofx
(11)
nosql
(3)
OO_Design
(1)
original_content
(4)
scriptUnixAutosys
(19)
SOA
(7)
socket/stream
(15)
sticky
(1)
subquery+join
(2)
swing
(32)
sybase
(6)
tech_orphan
(12)
tech+fin_career
(30)
telco
(11)
thread
(21)
timeSaver
(13)
tune
(10)
US_imm
(2)
US_misc
(2)
windoz
(20)
z_algo+dataStructure
(4)
z_arch
(2)
z_c#GUI
(30)
z_career
(10)
z_career]US^Asia
(2)
z_careerBig20
(1)
z_careerFinanceTech
(11)
z_FIX
(6)
z_forex
(31)
z_hib
(2)
z_ikm
(7)
z_inMemDB
(3)
z_j2ee
(10)
z_oq
(14)
z_php
(1)
z_py
(26)
z_quant
(4)
z_skillist
(3)
z_spr
(5)