Too small TCP TTLs

[Head UNIX Hacquer]

Problem: The time-to-live value for TCP/IP packets is too small to reach
	some hosts. 

Index:	/sys/netinet/tcp_timer.h  4.2 BSD, UMAX 3.X, SunOS 3.X

Description:
	Some hosts are reachable with ping(8) but cannot sustain TCP
	connections. 

Repeat-By:
	FTP to 'reagan.ai.mit.edu' from 'emx.utexas.edu'.

Fix:
	Increase TCP_TTL in /sys/netinet/tcp_timer.h to a MUCH larger value.

	The 4.2BSD ttl of 15 is not sufficent.  Many other systems with
	4.2BSD derived TCP/IP implimentations (SunOS 3.X, Encore UMAX 3.X,
	DYNIX 3.0.4 and Wollongong VMS TCP/IP at least) suffer from this.

	It is not clear that the 4.3BSD value of 30 will be sufficent
	in the future.  With the demise of ARPANET, sites
	that were only 2 or 3 hops away are now 6 or 10 hops away
	(such as the U of Texas).  Local routers (gateways) cause even more
	hops to be taken, requiring a longer TTL for a TCP packet to survive
	to where it needs to go.

Shouter-To-Dead-Parrots @ Univ. of Texas Computation Center; Austin, Texas  
	clyde at emx.utexas.edu; ...!cs.utexas.edu!ut-emx!clyde

Tip #268: Don't feel insecure or inferior! Remember, you're ORGANIC!!
	  You could win an argument with almost any rock!