[GNUnet-SVN] r1985 - GNUnet-docs/WWW

[grothoff]

Author: grothoff
Date: 2005-08-27 23:39:46 -0700 (Sat, 27 Aug 2005)
New Revision: 1985

Modified:
   GNUnet-docs/WWW/faq.php3
Log:
update

Modified: GNUnet-docs/WWW/faq.php3
===================================================================
--- GNUnet-docs/WWW/faq.php3    2005-08-28 04:08:06 UTC (rev 1984)
+++ GNUnet-docs/WWW/faq.php3    2005-08-28 06:39:46 UTC (rev 1985)
@@ -47,6 +47,7 @@
 W("GNUnet is free software, available under the %s (GPL).",
   extlink_("http://www.gnu.org/copyleft/gpl.html";,
            "GNU Public License"));
+W("You are free to run, distribute or modify the code under the terms stated 
in that license.");
 W("We are a part of the %s.",
   extlink_("http://www.gnu.org/","GNU project"));
 EP();
@@ -367,45 +368,22 @@
 W("The answer to this is, that encryption is incredibly fast.");
 W("GNUnet uses mostly AES-256, a very fast and secure cipher.");
 W("What really often makes anonymous file-sharing slow are artificial delays 
that were introduced to make timing analysis hard and to group messages into 
larger packets.");
-W("The best illustration for this is the following timing results obtained by 
inserting and downloading a 3 MB binary from the local host (no real networking 
involved, only loopback).");
-W("Insertion takes (on a Celeron 667, 400 MB RAM):");
-
-P();
-PRE("real    3m38.954s\n" .
-    "user    0m2.450s\n" .
-    "sys     0m0.410s");
-P();
-W("For the download, it takes:");
-P();
-PRE("real    0m13.796s\n" .
-    "user    0m2.150s\n" .
-    "sys     0m0.660s");
-
-P();
-W("As you can see, the <em>user</em> CPU usage (how much is actually used) is 
miniscule compared to the <em>real</em> time that it took to download the 
file.");
-W("The real time is longer because the harddrive could not perform the random 
accesses to the GNUnet databases any faster.");
-P();
-W("GNUnet thus uses CPU and network resources fairly responsibly.");
-W("The only thing that GNUnet &#8220;wastes&#8221; is time, in some sense, it 
sits idle.");
 W("The reason is, that this makes it harder to correlate actions.");
 W("GNUnet must wait for enough traffic from other peers to make it plausible 
that the traffic did not originate from the local peer.");
 W("Larger delays also allow for more reordering of messages by the individual 
peer.");
-W("GNUnet peers try to maximize bandwidth utility by reordering messages.");
 W("By allowing peers to delay messages, it is easier for them to build more 
efficient messages.");
-W("Once the load in GNUnet goes up, GNUnet will increase its actual 
utilization of resources up to the maximum specified by the user (see 
gnunetd.conf).");
-W("Thus, in some sense, GNUnet is slow but efficient.");
-
 P();
-
-W("The GNUnet developers are trying to make the perceived response times 
faster; yet, there is not much hope that the response times will get close to 
typical response times from other applications like the WWW.");
-W("Due to improvements in the code, the numbers quoted above may not always 
reflect the actual performance of the current version.");
-
+W("The primary cause of CPU consumption in the current implementation are 
algorithms for message scheduling.");
+W("GNUnet peers try to maximize bandwidth utility by reordering messages.");
+W("Also performing downloads in parallel can cause some significant accounting 
issues.");
+W("Many datastructures used currently are simple lists that take time linear 
to their size to operate on.");
+W("For local indexing operations the current release is typically pushing the 
limits of both the CPU and the harddrive.");
 P();
-
-W("Theoretically, it is possible that a download via GNUnet is even faster 
than a download from a crowded webserver or a single dialup user, but how 
likely this is depends in practice on the content -- and we neither promise nor 
really expect to achieve this level of performance.");
+W("We expect to use smarter, faster datastructures in the future to reduce CPU 
consumption.");
+W("The GNUnet developers are always trying to improve performance; yet, there 
is not much hope that performance will ever get close to typical response times 
from other applications like the WWW.");
+W("Theoretically, it is possible that a download via GNUnet is even faster 
than a download from a crowded webserver or a single dialup user, but how 
likely this is depends in practice on how the content is spread throughout the 
network -- and we neither promise nor really expect to achieve this level of 
performance.");
 W("While peer-to-peer networks can theoretically provide better performance 
than dedicated servers, their true strength lies in the possibility of being 
anarchistic: low administrative overhead, anonymity, no single point of 
failure.");
 W("Complete decentralization is very costly and we should thus not expect to 
outperform the centralized solution, especially not if we also want 
anonymity.");
-
 EP();
 
 ANCHOR("lastblocks"); H3("Why is downloading the last few blocks so slow?");