[ih] reinventing the wheel, was Internet History Lives on the Internet?

Richard Bennett richard at bennett.com
Tue Feb 26 13:24:21 PST 2019


Miles made the claim that BT is “more efficient” than unicast because it somehow becomes "EFFECTIVELY multicast - avoiding redundant traffic when more than one destination is downloading the same file, at the same time.” I don’t believe this is true. 

BT is a UDP-based protocol that senses queuing delays and defers more aggressively than TCP when they happen. Here’s the protocol spec: http://www.bittorrent.org/beps/bep_0003.html, if you can find any redundancy reduction in it please point it out.

RB

> On Feb 26, 2019, at 2:16 PM, Guy Almes <galmes at tamu.edu> wrote:
> 
> Richard,
>  If you do multicast at the network layer, you get IP multicast, Steve Deering's thesis project that escaped from the lab and has challenged network engineers ever since.
>  But if you do multicast at the application layer, you get lots of (IP) unicast traffic, but with some of the properties of IP multicast and also with some of the advantages of TCP and without the challenging routing issues of IP multicast.  Examples include BitTorrent, Unidata's LDM/IDD software, and the USENET netnews utility.
>  Oh, and you can also do multicast at the Ethernet level. Historically, it was the multicast feature of the then-new 10Mb/s Ethernet that triggered Deering's considering the possibility of multicast at the network layer.
> 
>  So IMHO the term multicast can be used appropriately without necessarily referring to network-layer IP multicast.
> 
>  Back to the primary topic,
> 	-- Guy
> 
> On 2/26/19 15:35, Richard Bennett wrote:
>> Perhaps you can explain that the term "EFFECTIVELY multicast” means to you, and how it applies to BitTorrent. Every time I’ve looked at it in Wireshark I see nothing but unicast.
>> RB
>>> On Feb 26, 2019, at 10:01 AM, Miles Fidelman <mfidelman at meetinghouse.net <mailto:mfidelman at meetinghouse.net>> wrote:
>>> 
>>> Well, that's not how it's supposed to work.  Or claimed to work.
>>> 
>>> On 2/25/19 8:27 PM, Richard Bennett wrote:
>>>> Naw, it’s all unicast with plenty of redundant traffic moving along long routes.
>>>> 
>>>>> On Feb 25, 2019, at 5:49 PM, Miles Fidelman <mfidelman at meetinghouse.net <mailto:mfidelman at meetinghouse.net>> wrote:
>>>>> 
>>>>> That's not what anybody said.  It's EFFECTIVELY multicast - avoiding redundant traffic when more than one destination is downloading the same file, at the same time.  That's its whole point.  (Or as someone else put it, it's a caching mechanism.)
>>>>> 
>>>>> Miles
>>>>> 
>>>>> On 2/25/19 4:24 PM, Richard Bennett wrote:
>>>>>> Bittorrent doesn’t use multicast.
>>>>>> 
>>>>>>> On Feb 25, 2019, at 1:42 PM, Miles Fidelman <mfidelman at meetinghouse.net <mailto:mfidelman at meetinghouse.net>> wrote:
>>>>>>> 
>>>>>>> Well yes, but torrent is a distribution mechanism - it's not an infrastructure for maintaining or mirroring files.  It's essentially another, sometimes more efficient, option for click-to-download (e.g., ftp, http, bittorrent).  And it's only more efficient if multiple people are downloading at the same time.
>>>>>>> 
>>>>>>> Miles
>>>>>>> 
>>>>>>> On 2/25/19 12:28 PM, Richard Bennett wrote:
>>>>>>>> No, Bittorrent uses DHTs and accesses as many copies of a given file as users choose to share. It has to be running to serve files, however.
>>>>>>>> 
>>>>>>>> RB
>>>>>>>> 
>>>>>>>>> On Feb 25, 2019, at 9:59 AM, Miles Fidelman <mfidelman at meetinghouse.net <mailto:mfidelman at meetinghouse.net>> wrote:
>>>>>>>>> 
>>>>>>>>> On 2/24/19 9:53 PM, John Levine wrote:
>>>>>>>>> 
>>>>>>>>>>> That's what I find intriguing about my Benevolent BotNet notion. Rather
>>>>>>>>>>> than depending on finding an institution interested in, competent at,
>>>>>>>>>>> and willing to save history, and hoping that it has longevity, you rely
>>>>>>>>>>> on a network of volunteers to provide that survivable infrastructure by
>>>>>>>>>>> volunteering their excess computing resources.
>>>>>>>>>> Hi again.  Please look at Bittorrent and tell us how it is different
>>>>>>>>>> from what you're proposing.
>>>>>>>>>> 
>>>>>>>>>> Bittorrent has the advantage of already existing and being deployed
>>>>>>>>>> all over the world.  It's notorious for pirated music but it's also
>>>>>>>>>> widely used for sharing linux distributions and the like.
>>>>>>>>>> 
>>>>>>>>> Bit torrent is transient.  It's more like an ad-hoc multi-cast
>>>>>>>>> streaming.  When nobody is downloading, there may be only one copy of
>>>>>>>>> the file.
>>>>>>>>> 
>>>>>>>>> Now gnutella, and some of the other P2P file sharing systems - that
>>>>>>>>> replicate copies, or distribute files across a distributed hash table -
>>>>>>>>> that's another story entirely.
>>>>>>>>> 
>>>>>>>>> -- 
>>>>>>>>> In theory, there is no difference between theory and practice.
>>>>>>>>> In practice, there is.  .... Yogi Berra
>>>>>>>>> 
>>>>>>>>> _______
>>>>>>>>> internet-history mailing list
>>>>>>>>> internet-history at postel.org <mailto:internet-history at postel.org>
>>>>>>>>> http://mailman.postel.org/mailman/listinfo/internet-history
>>>>>>>>> Contact list-owner at postel.org for assistance.
>>>>>>>> 
>>>>>>>>>>>>>>>> Richard Bennett
>>>>>>>> High Tech Forum <https://urldefense.proofpoint.com/v2/url?u=http-3A__hightechforum.org_&d=DwMFaQ&c=ODFT-G5SujMiGrKuoJJjVg&r=R09ttqQsnYaGPTsoYTZvLQ&m=L_GuN-mjLyb3nq7400xTV8D7FqWBXEnvfamqUrl6ys4&s=5sVVbAboWDOn4Em7VNOaohMIw5oSQNOpdgpOgTEb2co&e=> Founder
>>>>>>>> Ethernet & Wi-Fi standards co-creator
>>>>>>>> 
>>>>>>>> Internet Policy Consultant
>>>>>>>> 
>>>>>>> -- 
>>>>>>> In theory, there is no difference between theory and practice.
>>>>>>> In practice, there is.  .... Yogi Berra
>>>>>> 
>>>>>>>>>>>> Richard Bennett
>>>>>> High Tech Forum <https://urldefense.proofpoint.com/v2/url?u=http-3A__hightechforum.org_&d=DwMFaQ&c=ODFT-G5SujMiGrKuoJJjVg&r=R09ttqQsnYaGPTsoYTZvLQ&m=L_GuN-mjLyb3nq7400xTV8D7FqWBXEnvfamqUrl6ys4&s=5sVVbAboWDOn4Em7VNOaohMIw5oSQNOpdgpOgTEb2co&e=> Founder
>>>>>> Ethernet & Wi-Fi standards co-creator
>>>>>> 
>>>>>> Internet Policy Consultant
>>>>>> 
>>>>> -- 
>>>>> In theory, there is no difference between theory and practice.
>>>>> In practice, there is.  .... Yogi Berra
>>>>> _______
>>>>> internet-history mailing list
>>>>> internet-history at postel.org <mailto:internet-history at postel.org>
>>>>> http://mailman.postel.org/mailman/listinfo/internet-history
>>>>> Contact list-owner at postel.org for assistance.
>>>> 
>>>>>>>> Richard Bennett
>>>> High Tech Forum <https://urldefense.proofpoint.com/v2/url?u=http-3A__hightechforum.org_&d=DwMFaQ&c=ODFT-G5SujMiGrKuoJJjVg&r=R09ttqQsnYaGPTsoYTZvLQ&m=L_GuN-mjLyb3nq7400xTV8D7FqWBXEnvfamqUrl6ys4&s=5sVVbAboWDOn4Em7VNOaohMIw5oSQNOpdgpOgTEb2co&e=> Founder
>>>> Ethernet & Wi-Fi standards co-creator
>>>> 
>>>> Internet Policy Consultant
>>>> 
>>> -- 
>>> In theory, there is no difference between theory and practice.
>>> In practice, there is.  .... Yogi Berra
>>> _______
>>> internet-history mailing list
>>> internet-history at postel.org <mailto:internet-history at postel.org>
>>> http://mailman.postel.org/mailman/listinfo/internet-history
>>> Contact list-owner at postel.org for assistance.
>>>> Richard Bennett
>> High Tech Forum <https://urldefense.proofpoint.com/v2/url?u=http-3A__hightechforum.org&d=DwMFaQ&c=ODFT-G5SujMiGrKuoJJjVg&r=R09ttqQsnYaGPTsoYTZvLQ&m=L_GuN-mjLyb3nq7400xTV8D7FqWBXEnvfamqUrl6ys4&s=LZmy5fQZwLAb8wV6BIqZQ4PoPWOZ9QTG2meSKhwn4gc&e=> Founder
>> Ethernet & Wi-Fi standards co-creator
>> Internet Policy Consultant
>> _______
>> internet-history mailing list
>> internet-history at postel.org
>> http://mailman.postel.org/mailman/listinfo/internet-history
>> Contact list-owner at postel.org for assistance.

—
Richard Bennett
High Tech Forum <http://hightechforum.org/> Founder
Ethernet & Wi-Fi standards co-creator

Internet Policy Consultant

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20190226/fb80c176/attachment.htm>


More information about the Internet-history mailing list