Is Kazaa Nice
Violate copyrights?
Package semi-hidden extra software include ....
Change DNS server
Use CPU and disk for distributed
computing http://kazaa-download-kazaa.com/
Steal refered-to-seller commissions
Hates kazaalite
Kazaa goals
Make money through file sharing
Use peer-peer to save our bandwidth
and avoid legal problems
Show users adds to make revenue
Provide buy-through oppurtunites for
clients
Provide installs for other companies
add-ins
Kazaa is under attack
Legal problems
Kazaa would like to provide anonymous
downloads
Kazaa cannot exchange copyrighted
material
Hacker attacks
Kazaa lite and gIFT
Hacks attacking the network via custom
hacker software
RIAA can do this, as
well as script kidds
If Spammers could access
the banner ads, they just might
Data Attacks
RIAA could do this
Spammers could do this
Idiot Users
Who want to freeload
(download but not share)
Who change filenames
Network admins
Some just like firewalls
Some actively search out Kazaa
Some have legal and/or moral fears
Some have bandwidth concerns
1/2 to 1/2 of all
bandwidth on campuses is used by Kazaa
Joltid says that
between 50% and 80% of all bandwidth at ISPs is used by P-2-P
networks??? (see http://www.joltid.com/peercache.html)
Kazaa offers tools for
net admins to make Kazaa run more effeciently on their networks (saves
50% of P-2-P traffic)
Legal Stuff
Kazaa is HQed in Australia and encorperated in
Vanatua
Kazaa can be suid in the US
But what can a judge enforce?
Does Kazaa share too much?
http://kazaa-download-kazaa.com/techpaper1.htm
By the end of the
12 hour period 156 distinct users with shared inboxes were found.
To further demonstrate that this indicates unintentional file sharing,
we examined 20 distinct cases of shares on the inbox.dbx file by
manually using the "find more from same user" feature. 19 of the 20
users shared the other email files found in the default Microsoft
Outlook Express installation (Sent Items, Deleted Items, Outbox, etc.)
What's known about the achitecture?
There are no clients and servers
Everyone is both
Sometimes called servents
There are three levels of heirarchy
Normal servents
What you probably are
Supernodes
One servent per network
region elevates himself to be the supernode.
Every servent in the
region talks to the supernode
Each Supernode has an
index of all the servents's available files (see http://www.infoanarchy.org/wiki/wiki.pl?FastTrack)
Neat way for servents to
find local content before global content
Probably the central
server maybe doesn't have a complete list
How does one become a
supernode?
Why would one want to?
Central Servers
Each client has a list
compiled in
They are scattered
across national jurisdictions
They offer a list of
excluded-supernodes
What's known about the protocol?
Searches go to the SuperNodes, then to the Central
Servers
How to find the supernode
Each servent keeps a cache of ~200
supernodes
Candidates are chosen in the following
way
traceroute four IP
numbers (Scandanavia, North America, Japan, Australia)
Pick the common hops
(these represent your 'route to the world')
Look for a SuperNode on
the 'route to the world'
Probe them
Supernodes offer lists of other
supernodes
This probe uses UDP,then TCP as a
backup
Why UDP?
Why is there a TCP
backup?
Cache is stored in the registry at HKLM\KaZaA\ConnectionInfo\KazaaNet
There is no backup plan if they
failure to find a supernode???
That's hard to believe
Each SuperNode offers a 'busy' rating.
Connect to the client with the
smallest busyrating that's closest
Its based on HTTP and the GET command with region
modifiers
Easy to use, has lots of libraries one
can access
Relatively efficient
All GETs are done with a has value,
not a file name
Handles renames easily
Requires that all files
to be shared are read in their entirelty at least once
Is this
really done???
Easily entendable
It used to be accessable
As described in
Beginings of the Kazaa
handshake....http://www.blissed.org/~stucky/kazaa/handshake.html
http://www.infoanarchy.org/wiki/wiki.pl?FastTrack
Uses encryption to help against
hacker attack
What kind of
encryption???
How do they hide the
keys???
Maybe like
AOL and it's checksumming stuff
Each download is one or more segments (see http://kzfti.cjb.net/)
These segments represent the result of one http
get command
They are
stored in the *.dat file
They are
ordered by download time, not in file order
There is
also a 'status block' at the end of the *.dat file
Why the end??
When all
segments are downloaded, they are read off in file order to create the
file.
Hard for
them to do a 'partial file' viewer