System and method for protecting a TCP connection serving system from high-volume of TCP connection requests

1. A method of regulating TCP/IP connection requests which await service in a system by a TCP/IP connection control table to prevent overload thereof, said method comprising the steps of:

monitoring usage of said system on a dynamic basis,

based upon said usage, dynamically computing a time-out value T ho which defines the time duration that a TCP connection request may await service by said system,

removing from said TCP/IP connection control table all TCP/IP connection requests which have been awaiting service in said TCP/IP stack for a duration exceeding T ho; and

setting T ho =T min when N>N abs , when N>N limit setting T ho =max {T min , T′ ho /A}, where T′ ho is a previously existing value of T ho , where A>1, where N is the current usage of the table, and where 0≦N limit ≦N size , and when N≦N limit , setting T ho =min{T max , A*T′ ho /A};

wherein said TCP/IP connection control table has size N size and an upper bound for usable table size of N abs ≦N size , and where values of T ho are dynamically computed in a range [T min , T max ].

2. A method as set forth in claim 1 , comprising the steps of:

a) defining a plurality of table usage value N i spanning an increasing range of N i =0 to N i =N size ,

b) associating a corresponding plurality of time durations T i spanning a decreasing range of T i =T max to T i=T min , and

c) comparing current table usage N to N ho and setting T ho to a corresponding value T i .

3. A method as set forth in claim 1 , wherein T min has a value in a range of 0.01 to 1.0 secs. and wherein T max has a value in a range of 60 to 120 secs.