Clarification... I actually read the "content delivery
network<http://developer.yahoo.com/performance/rules.html#cdn>"
explanation, that wouldn't make sense for the typical university, but the
explanation found at http://developer.yahoo.com/performance/rules.html does
make sense for a portal installation which is what I was describing.

Parker

On Tue, May 13, 2008 at 10:03 AM, Parker Grimes <[EMAIL PROTECTED]> wrote:

> I think the "Use content delivery network" is a valid approach even for
> universities. At the very lease you can host your static content on a
> separate web server and take advantage of parallel downloads. We use this
> approach at SUU, all of our static content, CSS, images, JavaScript, are
> hosted on a different web server than our portal server.
>
> Yahoo's documentation says this:
>
> Splitting components allows you to maximize parallel downloads. Make sure
> you're using not more than 2-4 domains because of the DNS lookup penalty.
> For example, you can host your HTML and dynamic content on www.example.organd 
> split static components between
> static1.example.org and static2.example.org
>
> For more information check "Maximizing Parallel Downloads in the Carpool
> Lane <http://yuiblog.com/blog/2007/04/11/performance-research-part-4/>" by
> Tenni Theurer and Patty Chi.
> Parker
>
>
> On Tue, May 13, 2008 at 9:07 AM, Jen Bourey <[EMAIL PROTECTED]> wrote:
>
> > Comments inline . . .
> >
> > On Tue, May 13, 2008 at 10:53 AM, Eric Dalquist <
> > [EMAIL PROTECTED]> wrote:
> >
> > >  So from this report and Jen's email I think we could possibly look at
> > > the following features within the container.
> > >
> > > 1. Add a GZIP filter. This is generally done by folks in their HTTP
> > > server if they are interested though so I'm not sure it would be worth the
> > > effort on our part
> > >
> >
> > From the jQuery documentation, I'm not sure this is actually a good
> > idea.  Do we have a clear idea of the network download speed improvement vs.
> > client-side performance tradeoff?
> >
> >
> >
> > > 2. Add a filter that sets the expires header for static content. This
> > > would be great but we would need to address issues around file naming. 
> > > Using
> > > the expires header only really works if we're willing to never modify a 
> > > file
> > > without also changing its name. I'm not sure we are at a place where we 
> > > can
> > > try and dictate that from the framework without causing problems.
> > >
> >
> > Maybe we could address this by setting long-term headers for the
> > non-portal javascript, and leaving the portal's js as is?  It actually makes
> > quite a bit of sense to change the names of the jQuery main javascript and
> > plugins when we upgrade those to new versions.  The main jQuery file already
> > has the version number in its name, and the jQuery UI files probably would
> > benefit from having "-1.5beta" or something similar in their file names.
> >
> > I agree that setting long-term headers for the portal's javascript
> > probably isn't a great idea right now.  Since most of the javascript is
> > jQuery and other plugins though, we could probably set headers on 80% or so
> > of the javascript without making our lives painful.
> >
> >
> >
> > > 3. Add compile-time javascript minification. This probably won't be a
> > > huge gain but we can look into it as it should have minimal affect on
> > > uPortal implementers.
> > >
> >
> > Good plan.
> >
> >
> >  What may be more valuable is for someone to look into doing local
> > > tuning (setting up gzip and expires headers via apache) and documenting it
> > > as a tuning-recipe on the wiki.
> > >
> >
> > Sounds useful!
> >
> > - Jen
> >
> > --
> >
> > You are currently subscribed to [email protected] as: [EMAIL 
> > PROTECTED]
> >
> > To unsubscribe, change settings or access archives, see 
> > http://www.ja-sig.org/wiki/display/JSG/uportal-dev
> >
> >
>

-- 
You are currently subscribed to [email protected] as: [EMAIL 
PROTECTED]
To unsubscribe, change settings or access archives, see 
http://www.ja-sig.org/wiki/display/JSG/uportal-dev

Reply via email to