You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@httpd.apache.org by "Kamaraj, Jayakumar" <jk...@corp.untd.com> on 2009/11/17 09:56:56 UTC

[users@httpd] Just curious on SPDY

Hi,
            Just curious to know whether  Google announcement on SPDY
http://blog.chromium.org/2009/11/2x-faster-web.html needs change only in
Apache web server side or even needs change in application point of view
also. 
          Sorry to spam you guys .

Thanks,
Jai

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Re: [users@httpd] Just curious on SPDY

Posted by Mike Cardwell <ap...@lists.grepular.com>.
Nick Kew wrote:

>> I agree with the above. I started this thread to make people aware of
>> it's existance and to provoke discussion on the matter.
> 
> Yep, someone posted it a week or so back.

Yeah, me.

>>      However, if someone were to take up the reigns and begin
>> developing an Apache module for it using the open source code and
>> specs Google has published, I think the project has a more serious
>> chance of succeeding. I also think that an Apache with SPDY support
>> available before the spec is finalised would be in a stronger position
>> to influence how the protocol evolves during it's development.
> 
> Apache works mostly on a "scratch your own itch" basis, so that depends
> on whether anyone is sufficiently interested.
>
> Of course, if google is sufficiently interested, they have the
> capability to produce a module!

Yes, of course.

>> I also wonder if a transition like this to a new protocol could/should
>> be taken advantage of to get rid of the one SSL cert per IP:port
>> limitation we currently suffer from?
> 
> We dispensed with that limitation in (IIRC) 2.2.12.  See the change log.

As I understand it, you can't just stick multiple SSL sites on a single
IP:port atm without excluding all clients that don't support SNI. The
point I was trying to make was that something like SNI should be a
requirement in SPDY rather than just optional, otherwise we could end up
in the situation where some SPDY clients support SNI and others don't,
which I'm sure nobody wants.

-- 
Mike Cardwell - IT Consultant and LAMP developer
Cardwell IT Ltd. (UK Reg'd Company #06920226) http://cardwellit.com/
Technical Blog: https://secure.grepular.com/blog/

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Re: [users@httpd] Just curious on SPDY

Posted by Nick Kew <ni...@webthing.com>.
Mike Cardwell wrote:

> I agree with the above. I started this thread to make people aware of 
> it's existance and to provoke discussion on the matter.

Yep, someone posted it a week or so back.

>	 However, if 
> someone were to take up the reigns and begin developing an Apache module 
> for it using the open source code and specs Google has published, I 
> think the project has a more serious chance of succeeding. I also think 
> that an Apache with SPDY support available before the spec is finalised 
> would be in a stronger position to influence how the protocol evolves 
> during it's development.

Apache works mostly on a "scratch your own itch" basis, so that depends
on whether anyone is sufficiently interested.

Of course, if google is sufficiently interested, they have the
capability to produce a module!

> I also wonder if a transition like this to a new protocol could/should 
> be taken advantage of to get rid of the one SSL cert per IP:port 
> limitation we currently suffer from?

We dispensed with that limitation in (IIRC) 2.2.12.  See the change log.

-- 
Nick Kew

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Re: [users@httpd] Just curious on SPDY

Posted by Brian Mearns <me...@gmail.com>.
On Tue, Nov 17, 2009 at 9:55 AM, Mike Cardwell
<ap...@lists.grepular.com> wrote:
> Brian Mearns wrote:
>
>>>>           Just curious to know whether  Google announcement on SPDY
>>>> http://blog.chromium.org/2009/11/2x-faster-web.html needs change only in
>>>> Apache web server side or even needs change in application point of view
>>>> also.          Sorry to spam you guys .
>>>
>>> Both the server and the client would need to be updated in order to take
>>> advantage of it. If one or both don't support it, then the fallback would
>>> be
>>> normal HTTP.
>>>
>>> --
>>> Mike Cardwell - IT Consultant and LAMP developer
>>> Cardwell IT Ltd. (UK Reg'd Company #06920226) http://cardwellit.com/
>>> Technical Blog: https://secure.grepular.com/blog/
>>
>> [clip]
>>
>> Yes, SPDY is a new protocol which will require both the server and
>> client to support in order for it to work. However, from a user
>> perspective, I believe the goal is for it to be transparent. In other
>> words, if your browser and the web server it's talking to both support
>> SPDY, they will figure that out and use it. If either of them don't
>> support it, they'll just use plain old HTTP. Either way, you won't see
>> the difference as a user other than the potential speed benefits.
>>
>> Just to be clear, SPDY is far from being a new web-standard. Right
>> now, it's just a research project Google is undertaking: I think it's
>> going to be quite a while (a year at minimum) before any one (other
>> than Google, at least) thinks seriously about deploying it. But that's
>> just my $0.02.
>
> I agree with the above. I started this thread to make people aware of it's
> existance and to provoke discussion on the matter. However, if someone were
> to take up the reigns and begin developing an Apache module for it using the
> open source code and specs Google has published, I think the project has a
> more serious chance of succeeding. I also think that an Apache with SPDY
> support available before the spec is finalised would be in a stronger
> position to influence how the protocol evolves during it's development.

I understand your point, but I personally think it's too early in the
life of the spec to pull it from the sandbox. Putting it to actual use
in the wild before it's had a chance to mature at all will just cause
compatibility issues if and when the spec changes (which is likely
when it's such a young and relatively isolated thing, meaning it
hasn't had anybody from IETF or W3C or much of anybody else whack on
it at all).

>
> I also wonder if a transition like this to a new protocol could/should be
> taken advantage of to get rid of the one SSL cert per IP:port limitation we
> currently suffer from? Although the transition to ipv6 will get rid of this
> problem (lack of ip addresses) anyway without having to do any further work.

I really don't see how they're related. I think removing this
limitation is crucial if we're going to try to move towards a web that
requires SSL (as SPDY is currently slated for, I believe), but it
doesn't have anything to do with HTTP or SPDY, it's a limitation of
SSL itself. The SNI extension to SSL resolves the issue by essentially
allowing the equivalent of an HTTP Host: header to be included in the
SSL handshake. This is already supported in most modern web browsers,
and in Apache 2.2.12, I believe.

Cheers,
-Brian

>
> --
> Mike Cardwell - IT Consultant and LAMP developer
> Cardwell IT Ltd. (UK Reg'd Company #06920226) http://cardwellit.com/
> Technical Blog: https://secure.grepular.com/blog/
[snip]


-- 
Feel free to contact me using PGP Encryption:
Key Id: 0x3AA70848
Available from: http://keys.gnupg.net

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Re: [users@httpd] Just curious on SPDY

Posted by Mike Cardwell <ap...@lists.grepular.com>.
Brian Mearns wrote:

>>>            Just curious to know whether  Google announcement on SPDY
>>> http://blog.chromium.org/2009/11/2x-faster-web.html needs change only in
>>> Apache web server side or even needs change in application point of view
>>> also.          Sorry to spam you guys .
>> Both the server and the client would need to be updated in order to take
>> advantage of it. If one or both don't support it, then the fallback would be
>> normal HTTP.
>>
>> --
>> Mike Cardwell - IT Consultant and LAMP developer
>> Cardwell IT Ltd. (UK Reg'd Company #06920226) http://cardwellit.com/
>> Technical Blog: https://secure.grepular.com/blog/
> [clip]
> 
> Yes, SPDY is a new protocol which will require both the server and
> client to support in order for it to work. However, from a user
> perspective, I believe the goal is for it to be transparent. In other
> words, if your browser and the web server it's talking to both support
> SPDY, they will figure that out and use it. If either of them don't
> support it, they'll just use plain old HTTP. Either way, you won't see
> the difference as a user other than the potential speed benefits.
> 
> Just to be clear, SPDY is far from being a new web-standard. Right
> now, it's just a research project Google is undertaking: I think it's
> going to be quite a while (a year at minimum) before any one (other
> than Google, at least) thinks seriously about deploying it. But that's
> just my $0.02.

I agree with the above. I started this thread to make people aware of 
it's existance and to provoke discussion on the matter. However, if 
someone were to take up the reigns and begin developing an Apache module 
for it using the open source code and specs Google has published, I 
think the project has a more serious chance of succeeding. I also think 
that an Apache with SPDY support available before the spec is finalised 
would be in a stronger position to influence how the protocol evolves 
during it's development.

I also wonder if a transition like this to a new protocol could/should 
be taken advantage of to get rid of the one SSL cert per IP:port 
limitation we currently suffer from? Although the transition to ipv6 
will get rid of this problem (lack of ip addresses) anyway without 
having to do any further work.

-- 
Mike Cardwell - IT Consultant and LAMP developer
Cardwell IT Ltd. (UK Reg'd Company #06920226) http://cardwellit.com/
Technical Blog: https://secure.grepular.com/blog/

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Re: [users@httpd] Just curious on SPDY

Posted by Brian Mearns <me...@gmail.com>.
On Tue, Nov 17, 2009 at 6:15 AM, Mike Cardwell
<ap...@lists.grepular.com> wrote:
> Kamaraj, Jayakumar wrote:
>
>>            Just curious to know whether  Google announcement on SPDY
>> http://blog.chromium.org/2009/11/2x-faster-web.html needs change only in
>> Apache web server side or even needs change in application point of view
>> also.          Sorry to spam you guys .
>
> Both the server and the client would need to be updated in order to take
> advantage of it. If one or both don't support it, then the fallback would be
> normal HTTP.
>
> --
> Mike Cardwell - IT Consultant and LAMP developer
> Cardwell IT Ltd. (UK Reg'd Company #06920226) http://cardwellit.com/
> Technical Blog: https://secure.grepular.com/blog/
[clip]

Yes, SPDY is a new protocol which will require both the server and
client to support in order for it to work. However, from a user
perspective, I believe the goal is for it to be transparent. In other
words, if your browser and the web server it's talking to both support
SPDY, they will figure that out and use it. If either of them don't
support it, they'll just use plain old HTTP. Either way, you won't see
the difference as a user other than the potential speed benefits.

Just to be clear, SPDY is far from being a new web-standard. Right
now, it's just a research project Google is undertaking: I think it's
going to be quite a while (a year at minimum) before any one (other
than Google, at least) thinks seriously about deploying it. But that's
just my $0.02.

-Brian

-- 
Feel free to contact me using PGP Encryption:
Key Id: 0x3AA70848
Available from: http://keys.gnupg.net

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Re: [users@httpd] Just curious on SPDY

Posted by Mike Cardwell <ap...@lists.grepular.com>.
Kamaraj, Jayakumar wrote:

>             Just curious to know whether  Google announcement on SPDY
> http://blog.chromium.org/2009/11/2x-faster-web.html needs change only in
> Apache web server side or even needs change in application point of view
> also. 
>           Sorry to spam you guys .

Both the server and the client would need to be updated in order to take 
advantage of it. If one or both don't support it, then the fallback 
would be normal HTTP.

-- 
Mike Cardwell - IT Consultant and LAMP developer
Cardwell IT Ltd. (UK Reg'd Company #06920226) http://cardwellit.com/
Technical Blog: https://secure.grepular.com/blog/

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org