[ENet-discuss] Enet Header size

Daniel Aquino mr.danielaquino at gmail.com
Wed Jan 7 15:46:02 PST 2009


There is also an ability to flush based on a certain peer I believe...

I would think that certain real time data would be desirable to flush right
away to the peer...

Then remaining packets could be picked up at some interval...

On Wed, Jan 7, 2009 at 5:47 PM, Nuno Silva <little.coding.fox at gmail.com>wrote:

> On the docs it said that enet_host_flush() is also supposed to send all
> queued packets (for sending), although it wont treat events like
> enet_host_flush(), so if the OP just wants to send various packets and not
> wait for enet_host_service(), he could use enet_host_flush() instead.
>
>
> On Thu, Jan 8, 2009 at 6:29 AM, Lee Salzman <lsalzman1 at cox.net> wrote:
>
>> This all depends on what is acceptable latency for you and how often you
>> are sending packets. Packets are sent out when you call enet_host_service(),
>> so schedule this according to your needs.
>>
>> Lee
>>
>>
>> Peter Soxberger wrote:
>>
>>> In what periods should enet_host_service() be called? I'm using a thread
>>> and alwas wait 2 milliseconds until I call enet_host_service(). Should I
>>> wait longer? What are your typical periods?
>>>
>>> Greets,
>>> Peter Soxberger
>>>
>>> -------- Original-Nachricht --------
>>>
>>>
>>>> Datum: Wed, 07 Jan 2009 20:11:06 -0800
>>>> Von: Lee Salzman <lsalzman1 at cox.net>
>>>> An: Discussion of the ENet library <enet-discuss at cubik.org>
>>>> Betreff: Re: [ENet-discuss] Enet Header size
>>>>
>>>>
>>>
>>>
>>>
>>>> It collects packets in between calls to enet_host_service(), yes.
>>>>
>>>> Lee
>>>>
>>>> Peter Soxberger wrote:
>>>>
>>>>
>>>>> Ok does this mean that ENet collects the small packets and send it as
>>>>>
>>>>>
>>>> one big packet for saving overhead?
>>>>
>>>>
>>>>> I worried that this isn't implented because I use a lot of packets with
>>>>>
>>>>>
>>>> small data in them and this would produce a lot of unnecessary overhead.
>>>>
>>>>
>>>>> -------- Original-Nachricht --------
>>>>>
>>>>>
>>>>>> Datum: Wed, 07 Jan 2009 19:18:15 -0800
>>>>>> Von: Lee Salzman <lsalzman1 at cox.net>
>>>>>> An: Discussion of the ENet library <enet-discuss at cubik.org>
>>>>>> Betreff: Re: [ENet-discuss] Enet Header size
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>>> The size of a header for an individual user packet WITHIN a protocol
>>>>>> packet is about 6-8 bytes. Each protocol packet (which groups as many user
>>>>>> packets as it can up to the MTU) has a header size of 8 bytes.
>>>>>>
>>>>>> Lee
>>>>>>
>>>>>> Peter Soxberger wrote:
>>>>>>
>>>>>>
>>>>>>> Hi!
>>>>>>>
>>>>>>> I'm currently improving my code that calculates the actual
>>>>>>>
>>>>>>>
>>>>>> transferrate.
>>>>
>>>>
>>>>>
>>>>>>>
>>>>>> Does anybody of you know, how much overhead I produce if I send a
>>>>>>
>>>>>>
>>>>> single
>>>>
>>>>
>>>>> packet? I know that there is the UDP Header which is 8 Byte big. But
>>>>>>
>>>>>>
>>>>> how
>>>>
>>>>
>>>>> much is used by ENet?
>>>>>>
>>>>>>
>>>>>>> If I calculate the real size of a packet, is this correct:
>>>>>>>
>>>>>>> [UDP Header]+[ENet Headers]+[packet->dataLength] = Packet Size
>>>>>>>
>>>>>>> Thanks for your help!
>>>>>>>
>>>>>>>
>>>>>> _______________________________________________
>>>>>> ENet-discuss mailing list
>>>>>> ENet-discuss at cubik.org
>>>>>> http://lists.cubik.org/mailman/listinfo/enet-discuss
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>> _______________________________________________
>>>> ENet-discuss mailing list
>>>> ENet-discuss at cubik.org
>>>> http://lists.cubik.org/mailman/listinfo/enet-discuss
>>>>
>>>>
>>>
>>>
>>>
>>
>> _______________________________________________
>> ENet-discuss mailing list
>> ENet-discuss at cubik.org
>> http://lists.cubik.org/mailman/listinfo/enet-discuss
>>
>
>
> _______________________________________________
> ENet-discuss mailing list
> ENet-discuss at cubik.org
> http://lists.cubik.org/mailman/listinfo/enet-discuss
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cubik.org/pipermail/enet-discuss/attachments/20090107/4422e72c/attachment-0001.htm>


More information about the ENet-discuss mailing list