Search String: Display: Description: Sort:

Results:

References: [ +subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*TOE\s+brain\s+dump\s*$/: 106 ]

Total 106 documents matching your query.

41. Re: TOE brain dump (score: 1)
Author: xxxxxx>
Date: Wed, 6 Aug 2003 02:13:04 -0300
Hmm, but doesn't it also unify transport semantics (i.e. chop TCP streams into messages), maybe add reliability to transports that don't have it, and provide addressing ? Okay, perhaps you wouldn't c
/archives/netdev/2003-08/msg00152.html (11,772 bytes)

42. Re: TOE brain dump (score: 1)
Author: xxxxxx>
Date: Wed, 6 Aug 2003 00:12:36 -0700 (PDT)
Do be sure to check that your data payload is correct. Everyone knows that a router/gateway/switch with a sticky bit in its memory will recompute the net crc16 checksum insure it pass the to the nic
/archives/netdev/2003-08/msg00154.html (14,966 bytes)

43. Re: TOE brain dump (score: 1)
Author: xxxxxx>
Date: 06 Aug 2003 01:58:56 -0600
Those are all features of the MPI implementation. It is not that MPI does not have an underlying transport. MPI has a lot of underlying transports. And there is a different MPI implementation for eac
/archives/netdev/2003-08/msg00158.html (12,817 bytes)

44. Re: TOE brain dump (score: 1)
Author: xxxxxx>
Date: Wed, 06 Aug 2003 18:20:06 +1000
At 05:12 PM 6/08/2003, Andre Hedrick wrote: Do be sure to check that your data payload is correct. Everyone knows that a router/gateway/switch with a sticky bit in its memory will recompute the net c
/archives/netdev/2003-08/msg00160.html (10,594 bytes)

45. Re: TOE brain dump (score: 1)
Author: xxxxxx>
Date: Wed, 6 Aug 2003 01:22:18 -0700
To make sure emails about Falun Gong and other undesirable topics don't make it into China.
/archives/netdev/2003-08/msg00161.html (9,717 bytes)

46. Re: TOE brain dump (score: 1)
Author: val@xxxxxxxxxx>
Date: Wed, 6 Aug 2003 07:46:33 -0500
I think you are off here since the second and third layer should not recompute checksums other than for the header (if they even did that). Most of the switches I used (mind, not configured) were wir
/archives/netdev/2003-08/msg00166.html (14,183 bytes)

47. Re: TOE brain dump (score: 1)
Author: xxxxxxxxxxxxxx>
Date: Wed, 6 Aug 2003 08:07:37 -0500
Thats not a router, or switch... It's a firewall :-)
/archives/netdev/2003-08/msg00167.html (10,000 bytes)

48. Re: TOE brain dump (score: 1)
Author: xxxxxxxxxxxxxx>
Date: Wed, 6 Aug 2003 10:37:58 -0300
Ah, the joy of herding cats :-) But I guess you just need a sufficiently competent and sufficiently well-funded group that goes ahead and does it. There is usually little point in directly involving
/archives/netdev/2003-08/msg00168.html (10,943 bytes)

49. Re: TOE brain dump (score: 1)
Author: xxxxxxxxxxxxxx>
Date: Wed, 6 Aug 2003 11:25:56 -0500
The switches may be "wire speed" but that doesn't help the latency any. AFAIK all GigE switches are store-and-forward, which automatically costs you the full 1.3us for each link hop. (I didn't check
/archives/netdev/2003-08/msg00169.html (13,430 bytes)

50. Re: TOE brain dump (score: 1)
Author: on@xxxxxxxxxxx>
Date: Wed, 6 Aug 2003 13:58:59 -0500
If it reads the packet into memory before starting transmission, it isn't "wire speed". It is a router. Just because the packet gets transfered to a buffer correctly does not mean that buffer is the
/archives/netdev/2003-08/msg00179.html (15,635 bytes)

51. Re: TOE brain dump (score: 1)
Author: t@xxxxxxxxxxxxx
Date: Wed, 6 Aug 2003 14:39:56 -0500
[Please read an implied "I might be totally off base here, since I've never designed an Ethernet switch" disclaimer into this paragraph.] This statement is completely false. Ethernet switches *do* re
/archives/netdev/2003-08/msg00181.html (17,827 bytes)

52. Re: TOE brain dump (score: 1)
Author: .hen@xxxxxxxxx>
Date: Tue, 5 Aug 2003 10:15:25 +0200
No, since I didn't state that a retransmission is done along the pipe, because you cannot go back in a pipeline. A retransmission can be done at the end of the pipe, where this can also be done in ha
/archives/netdev/2003-08/msg00195.html (9,922 bytes)

53. RE: TOE brain dump (score: 1)
Author: xxxxxxxxxxxxxx>
Date: Wed, 6 Aug 2003 14:13:47 -0700
Some do. Some don't. Some are configurable. If they use cut-through switching, they do. Some use adaptive switching, which means they use cut-through switching but change to store and forward if the
/archives/netdev/2003-08/msg00196.html (10,470 bytes)

54. TOE brain dump (score: 1)
Author: Werner Almesberger <werner@xxxxxxxxxxxxxxx>
Date: Sat, 2 Aug 2003 14:04:44 -0300
At OLS, there was a bit of discussion on (true and false *) TOEs (TCP Offload Engines). In the course of this discussion, I've suggested what might be a novel approach, so in case this is a good idea
/archives/netdev/2003-08/msg01009.html (13,617 bytes)

55. Re: TOE brain dump (score: 1)
Author: Nivedita Singhvi <niv@xxxxxxxxxx>
Date: Sat, 02 Aug 2003 10:32:55 -0700
(*) The InfiniBand people unfortunately call also their TCP/IP bypass "TOE" (for which they promptly get shouted down, every time they use that word). This is misleading, because Thank you! Yes! All
/archives/netdev/2003-08/msg01010.html (10,701 bytes)

56. Re: TOE brain dump (score: 1)
Author: Werner Almesberger <werner@xxxxxxxxxxxxxxx>
Date: Sat, 2 Aug 2003 15:06:00 -0300
I have some hope that sendfile plus a NUMA-like approach will be sufficient for keeping transfers away from buses and memory they don't need to touch. Many companies default to "closed" designs if th
/archives/netdev/2003-08/msg01011.html (10,702 bytes)

57. Re: TOE brain dump (score: 1)
Author: Jeff Garzik <jgarzik@xxxxxxxxx>
Date: Sat, 02 Aug 2003 15:08:52 -0400
My own brain dump: If one wants to go straight from disk to network, why is anyone bothering to involve the host CPU and host memory bus at all? Memory bandwidth and PCI bus bandwidth are still bottl
/archives/netdev/2003-08/msg01012.html (11,103 bytes)

58. Re: TOE brain dump (score: 1)
Author: Alan Cox <alan@xxxxxxxxxxxxxxxxxxx>
Date: 02 Aug 2003 21:57:44 +0100
or even the protocol and protocol refinements.. Like say an opteron in the 2nd socket on the motherboard It moves the cost it doesnt make it vanish If I read you right you are arguing for a second pr
/archives/netdev/2003-08/msg01013.html (10,057 bytes)

59. Re: TOE brain dump (score: 1)
Author: Werner Almesberger <werner@xxxxxxxxxxxxxxx>
Date: Sat, 2 Aug 2003 18:49:01 -0300
But then, ten "optimized xxPs" that work well in two different scenarios each, but not so good in the 98 others, wouldn't be much fun either. It's been tried a number of times. Usually, real life sne
/archives/netdev/2003-08/msg01014.html (11,325 bytes)

60. Re: TOE brain dump (score: 1)
Author: Werner Almesberger <werner@xxxxxxxxxxxxxxx>
Date: Sat, 2 Aug 2003 19:14:11 -0300
I don't think it really can. What it can do is reduce the overhead (which usually translates to latency and burstiness) and the sharing. That's taking this idea to an extreme, yes. I'd think of using
/archives/netdev/2003-08/msg01015.html (10,224 bytes)


This search system is powered by Namazu