Packeteer ... How it works

Without looking at content, it looks at traffic, and application information
at layer 7 * well below most equipments capabilities...
"Typically non-business in nature, bandwidth-intensive P2P applications
have presented legal and security challenges to business organizations.
Although business-critical P2P applications do exist, each application has
its own bandwidth habits and performance requirements. Therefore, each needs
its own appropriate performance management. A good performance management
strategy requires the ability to identify P2P applications and control their
bandwidth utilization.

Packeteer's application traffic management system identifies P2P
applications and helps you define appropriate policy management strategies.
These strategies can range from a total application block, to a small
trickle of bandwidth, to immediate and first access to bandwidth. At the
same time, the Packeteer system protects the business applications that
share the network and its resources with P2P traffic.

Packeteer's Layer 7 traffic classification identifies applications running
across your network - both business and non-business in nature. Whether
they've tunnelled through HTTP tunnel gateways or not, the Packeteer system
discovers and tracks P2P applications like KaZaA, Morpheus, Gnutella, iMesh,
and AudioGalaxy. It is this Layer 7 monitoring that enables Packeteer to
counter P2P's port-hopping tendencies, distinguishing it from ill-equipped
routers and firewalls. Once P2P traffic is identified, Packeteer applies
decisive policy controls that eliminate or minimize P2P's presence on the
network. You can block P2P, contain it to a reasonable bits-per-second rate,
and/or limit each user to a maximum. You can even discourage usage and avoid
blocking by providing P2P with such a small trickle of bandwidth that users
experience dismal performance and refrain from using the applications
completely. While the Packeteer system controls P2P traffic, it allocates
appropriate amounts of bandwidth to business applications to ensure
efficient and reliable performance."

The listing of P2P applications effectively blocked is pretty comprehensive

http://www.packeteer.com/resources/p...nDiscovery.pdf

As its primarily a blocker for corporate networks, I couldn't comment on how
it scales for full ISP's but I recall a university in the UK installing it,
and seeing a 90% reduction in traffic - pre Bittorrent...

In MsgID<BDCFA755.2E1B5%(E-Mail Removed)> within
uk.telecom.broadband, 'Graham in Melton' wrote:

> "Typically non-business in nature, bandwidth-intensive P2P applications
>have presented legal and security challenges to business organizations.
>Although business-critical P2P applications do exist, each application has
>its own bandwidth habits and performance requirements. Therefore, each needs
>its own appropriate performance management. A good performance management
>strategy requires the ability to identify P2P applications and control their
>bandwidth utilization.

Are they ever going to stop categorising traffic as "P2P"
- Almost all traffic on the internet is peer to peer..

If you blocked 'all p2p traffic' then you could certainly forget about
web browsing, ftp, email etc.

I wish the more widespread communal file sharing apps had been written
around port 21 or port 80 and had used a standard protocol, that would
have spoilt the saboteurs wordplay..

Where home use is concerned it would have been funny to watch the
other variety of saboteur trying to outlaw the running of a http or
ftp server..

--
Dave Johnson - (E-Mail Removed)

On Sun, 28 Nov 2004 23:05:24 UTC, Dave J <(E-Mail Removed)> wrote:

> Are they ever going to stop categorising traffic as "P2P"
> - Almost all traffic on the internet is peer to peer..
>
> If you blocked 'all p2p traffic' then you could certainly forget about
> web browsing, ftp, email etc.

Those three are are client-server, not peer-to-peer.

--
Bob Eager
begin a new life...dump Windows!

In MsgID<176uZD2KcidF-pn2-(E-Mail Removed)> within
uk.telecom.broadband, 'Bob Eager' wrote:

>> If you blocked 'all p2p traffic' then you could certainly forget about
>> web browsing, ftp, email etc.
>
>Those three are are client-server, not peer-to-peer.

All hosts on the internet are peers, all hosts have the same interface
to the network, IIRC it says so explicitly in an RFC somewhere, it
certainly assumes such throughout.

If you think about it, what is the difference between an http service
and a gnutella style service? You might initially say direction of
file transfer, whereupon I'd point out that http can receive files
quite nicely.

Suppose I wrote a communal file-share app that worked around http?
Relatively easy to do, would you describe it as "p2p"? - I would.

--
Dave Johnson - (E-Mail Removed)

On Mon, 29 Nov 2004 11:13:39 UTC, Dave J <(E-Mail Removed)> wrote:

> >Those three are are client-server, not peer-to-peer.
>
> All hosts on the internet are peers, all hosts have the same interface
> to the network, IIRC it says so explicitly in an RFC somewhere, it
> certainly assumes such throughout.
>
> If you think about it, what is the difference between an http service
> and a gnutella style service? You might initially say direction of
> file transfer, whereupon I'd point out that http can receive files
> quite nicely.
>
> Suppose I wrote a communal file-share app that worked around http?
> Relatively easy to do, would you describe it as "p2p"? - I would.

So would I. None of the nodes involved can be viewed as a server, at
least all of the time.

I'm talking roles, not protocols...there is a definite server/client
relationship in mail, web, etc. Not true with (say) gnutella. Hence the
latter is peer to peer, and the former aren't.

But we could argue over semantics for ages....!

--
Bob Eager
begin a new life...dump Windows!

In MsgID<176uZD2KcidF-pn2-(E-Mail Removed)> within
uk.telecom.broadband, 'Bob Eager' wrote:

>But we could argue over semantics for ages....!

I'm sorry to say that I would probably enjoy it.

I have a lifelong love of the English language, and I enjoy attempting
precision in the things that I say. Thanks for the playtime.

Dave J.

Dave J wrote:

> In MsgID<176uZD2KcidF-pn2-(E-Mail Removed)> within
> uk.telecom.broadband, 'Bob Eager' wrote:
>
> >But we could argue over semantics for ages....!
>
> I'm sorry to say that I would probably enjoy it.
>
> I have a lifelong love of the English language, and I enjoy attempting
> precision in the things that I say. Thanks for the playtime.
>

Well I've always considered myself crap at English.

The word peer to me implies that there is something in someway equal to me
ie. my kids can't get peer pressure from me.

When used in P2P peer means that I am connecting to another "end user" like
myself - indeed I may chat to them on a one to one basis. I can't compare
myself with a server that has loads more bandwidth and a one to many
relationship with clients.

Andy.

In MsgID<j5eu72-(E-Mail Removed)> within uk.telecom.broadband,
'Andy Furniss' wrote:

>> I have a lifelong love of the English language, and I enjoy attempting
>> precision in the things that I say. Thanks for the playtime.
>>
>
>Well I've always considered myself crap at English.

Heh - I didn't say that I was good, merely interested

>
>The word peer to me implies that there is something in someway equal to me
>ie. my kids can't get peer pressure from me.

Yeah, that was almost my point. On 'the' internet all hosts are equal,
they all have the same interface and once a connection has been made
there is no difference between the two ends.

Much documentation that I've read refers to 'peers' when talking about
client-server connections. As I said before, I'm fairly sure it's
actually stated in one of the defining RFCs, though I could easily be
wrong there.


>When used in P2P peer means that I am connecting to another "end user" like
>myself - indeed I may chat to them on a one to one basis.

>I can't compare
>myself with a server that has loads more bandwidth and a one to many
>relationship with clients.

And yet I often run an ftp server and sometimes a web server on this
machine. I'm only connected by pstn dialup..

--
Dave Johnson - (E-Mail Removed)

Dave J Wrote:
> I wish the more widespread communal file sharing apps had been written
> around port 21 or port 80 and had used a standard protocol, that would
> have spoilt the saboteurs wordplay..

i am a principle developer of Rod
http://larytet.sourceforge.net/btRat.shtml
Part of the functional requirements of Rodi is packet obfuscation.

i invite beta testers to run the application

arkad

--
larytet