Fast forward to 2009. Yahoo is now in the process of opening up Traffic Server as an Apache project. It’s already in incubator. Yahoo says it’s capable of 30,000 requests per server. Noteworthy is that this runs on generic hardware.

These days most websites use either Squid, Nginx, Pound or Varinish on the open source side. On the proprietary side there’s Citrix NetScaler, Foundry (now Brocade) ServerIron, Zeus ZXTM or F5’s Big-IP. The proprietary side can be either expensive software running on generic hardware or an appliance (which is generally a Intel based server with a custom modified Linux install for low maintenance and top performance).

At this point it’s apparently not 64-bit and doesn’t have native IPv6 support. However it appears to be usable and likely competitive with some of the other stuff out there already. Yahoo has been using it all along, and I hear they are pretty popular (problems aside).

It should be noted that commercial CDN’s aren’t really an alternative for reverse proxy or load balancer since they still require a robust and redundant origin. If anything they will reduce your requirements, not eliminate them.

Given everyone’s interest in scaling computing quickly and cheaply this is pretty noteworthy open source event. It tends to be an afterthought but these applications can be critical. Squid handles 78% of Wikipedia’s requests. Given all their traffic, you can see how it matters.

It will be interesting to see if a community builds around Traffic Server and if it sees adoption.

It has now relaunched using open source Drupal. Also interesting is that it’s no longer broadcasting any headers regarding it’s server. Considering Drupal is by far better tested on a Unix OS andApache, I’m wondering if they dropped Windows Server/IIS 6.0 in favor of some sort of Linux and Apache. I can’t find any hint at what they are using.

It’s noteworthy that Drupal was already used on recovery.gov and has been used in politics by way of CivicSpace for the Dean campaign in 2004.

Via Drupal it’s still using jQuery (verison 1.2.6). It’s also now using RSS rather than ATOM for feeds, which I presume is by way of the switch to Drupal rather than an intentional effort.

Another interesting change is they tweaked the doctype from XHTML Transitional to XHTML+RDFa.

Pretty much everything else is still the same including the design. Analytics is still done using WebTrends (holdover from the Bush administration) and Akamai still sits in front of their servers.

For CSS hackers: They still choose conditional CSS for IE compatibility.

Their pages don’t fully validate anymore, though there is no terrible markup either.

Video is still done using Flash, maybe they’ll consider adopting HTML5 video. They could do so and fallback to Flash. The latest versions of Firefox, Safari, and Chrome could take advantage of it today. The rest of the browsers would get the Flash experience. That would be the next major step in opening up. Mark Pilgrim has a good primer if they need.

Edit [9/26/2009 @ 1:45 PM EST]: Tim O’Reilly confirms it is indeed running on LAMP, specifically Red Hat Linux with Apache, MySQL and obviously PHP. Apache Solr is used for search.

I had heard about the ongoing NetApp vs. Sun patent war where NetApp feels that ZFS is too close to WAFL. It seems likely that Apple doesn’t want to get involved in that. Apple even has a fear of that potential with OGG Theora. Once the transition was made to ZFS it would be a costly and time-consuming effort to swap with something else since Mac OS has never been very file system neutral.

A new theory is that the Oracle/Sun deal leaves the company developing two filesystems: ZFS and Btfs. It sounds like Oracle’s Btfs is the more likely future. If Apple switched to ZFS they would have been left as the only platform using it. Linux can’t fully switch since the CDDL license isn’t fully compatible with GPL meaning it would need to be implemented through FUSE. Btfs is coming along for Linux.

Reading through Btfs, it seems like a lot of the big advantages of ZFS are already in Btfs though it lacks full disk encryption. It does however add online resizing. It’s also GPL and has support from RedHat, Novell, IBM and was accepted into the Linux mainline kernel as of 2.6.29rc1. That means it already has a much more robust community and seems likely to be widely accepted in UNIX land.

So will Apple switch to Btfs for Mac OS X 10.7 or 10.8? I think the two possibilities are that it will either build something in-house, or switch to Btfs. I think Btfs offers a compelling set of features and would allow Apple to brag about more compatibility with other OS’s and potentially adopt features as the file system matures at a low-cost. It’s possible we’ll hear something as soon as WWDC 2010.

Edit [10/25/2009 @ 6:30 PM EST]: Confirmed via Daring Fireball

Nobody mentioned there’s a removespikes.pl script (download) that will remove these outliers from your rrds. I put together a quick shell script to make it quick for when I need to run it again:

!/bin/sh

for i in /path/to/graphs/rrd/*;
do
        perl removespikes.pl $i;
done;

If you have a ton of graphs a quick shell script to iterate through the directly may be quicker. If you only have a handful like me, no big deal.

Keep the script around for the next time you have spikes to deal with.

Since these improvements are in the encoder rather than the format, or the playback library that means existing Theora users, as well as all Firefox 3.5 users will be benefiting from the work already done, as well as work done in the future without needed continued software updates, though I bet even playback will get some improvements over time.

Even better is that open video is free unlike most other formats out there.

I wish all browser vendors could agree and sync engines a bit more so that Safari/Chrome would ship the same version of WebKit rather than stagger based on their own release schedule. Same for Mozilla/Flock etc. While very difficult in many respects it would make it much easier on web developers to have less products out there to test against. I think it’s unlikely to happen, but that doesn’t mean it wouldn’t be convenient. Some Web Developer bias speaking here.

[Via: Tony Chang]

I’m thrilled to see a push for open video. Better encoders and decoders along with working with the Wikimedia Foundation (Wikipedia’s use of Theora can be very influential) will hopefully provide a boost for these formats which tends to be a cyclical trend once it gains momentum.

Unfortunately the build directions for Mac OS X are lacking (they don’t really exist). Building the dependencies didn’t really work for me. For anyone who wants to follow this, note that these are changes to the instructions, not replacements. It doesn’t discuss everything. I have only done this on Mac OS X 10.4.11. I have no idea if this works on any other version, though I suspect 10.5 will work fine.

For starters I used the following LDFLAGS:

LDFLAGS="-Wl,-blibpath:${INSTALL_DIR}/lib"

I’m not sure if -Wl is necessary or not, but I don’t think it’s harmful.

In general the directions listed on the build page are pretty strait forward and work fine on Mac OS X but I did encounter an issue once I started to compile Pango 1.17.5:

 gcc -DHAVE_CONFIG_H -I. -I.. -DG_LOG_DOMAIN=\"Pango\" -DPANGO_ENABLE_BACKEND -DPANGO_ENABLE_ENGINE -DSYSCONFDIR=\"/usr/local/rrdtool-1.3.5/etc\" -DLIBDIR=\"/usr/local/rrdtool-1.3.5/lib\" -I.. -DPANGO_ENABLE_DEBUG -D_REENTRANT -I/usr/local/rrdtool-1.3.5/include/glib-2.0 -I/usr/local/rrdtool-1.3.5/lib/glib-2.0/include -I/usr/local/rrdtool-1.3.5/include/cairo -I/usr/local/rrdtool-1.3.5/include/freetype2 -I/usr/local/rrdtool-1.3.5/include -I/usr/local/rrdtool-1.3.5/include/libpng12 -I/usr/local/rrdtool-1.3.5/include/pixman-1 -I/usr/local/rrdtool-1.3.5/include -I/usr/local/rrdtool-1.3.5/include/freetype2 -I/usr/local/rrdtool-1.3.5/include -xobjective-c -O3 -fPIC -Wall -MT libpangocairo_1_0_la-pangocairo-fontmap.lo -MD -MP -MF .deps/libpangocairo_1_0_la-pangocairo-fontmap.Tpo -c pangocairo-fontmap.c  -fno-common -DPIC -o .libs/libpangocairo_1_0_la-pangocairo-fontmap.o
In file included from pangocairo-fontmap.c:29:
pangocairo-atsui.h:27:25: error: cairo-atsui.h: No such file or directory
make[4]: *** [libpangocairo_1_0_la-pangocairo-fontmap.lo] Error 1
make[3]: *** [all-recursive] Error 1
make[2]: *** [all] Error 2
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

After some experimentation I found that I had to update to a newer version of Pango. Which needs a newer version of GLib and Cairo. Which requires a newer version of Pixman. You’ve got to love how that cascades. Here’s what I changed:

Pixman

I ended up using Pixman 0.13.2 and doing:

tar -xzf pixman-0.13.2.tar.gz
cd pixman-0.13.2
./configure --prefix=$INSTALL_DIR CFLAGS="-O3 -fPIC"
make
make install

Cairo

Then I installed Cairo 1.8.6 using:

tar -xzf cairo-1.8.6.tar.gz
cd cairo-1.8.6
./configure --prefix=$INSTALL_DIR --enable-xlib=no --enable-xlib-xrender=no --enable-win32=no  --enable-quartz=yes --enable-quartz-font=yes CFLAGS="-O3 -fPIC"
make
make install

I modified a few configure flags to use quartz though I left --enable-quartz-image off because it’s experimental as of 1.8.6. Also some of the flags changed slightly between versions (xlib-render is now xlib-xrender).

GLib

I then upgraded GLib to 2.18:

tar -xzf glib-2.18.tar.gz
cd glib-2.18
./configure --prefix=$INSTALL_DIR CFLAGS="-O3 -fPIC"
make
make install

This took forever to build, but it worked fine.

Pango

I then finally installed Pango 1.22.4:

tar -xzf pango-1.22.4.tar.gz
cd pango-1.22.4
./configure --prefix=$INSTALL_DIR --disable-tcl --disable-python
make
make install

If for some reason Pango doesn’t include Cairo as it’s backend like this when ./configure ends:

configuration:
        backends: Cairo FreeType X

Try using the following statements first:

CAIRO_CFLAGS="-I=/usr/local/rrdtool-1.3.5/include/cairo"
CAIRO_LIBS=/usr/local/rrdtool-1.3.5/lib/

Then run ./configure again.

Then compile RRDtool itself by following the directions as normal. This worked fine for me.

Cascading dependencies suck. There may have been an easier fix to my problem, but I wasn’t able to find it. Simply updating to newer versions of these libraries seemed like the best solution. That said, these aren’t officially sanctioned versions by the RRDtool team. Be warned.

Good luck to anyone else who tries this.

The machine is as thin as possible, runs low end hardware and has a single button for powering it on and off, headphone jacks, a built in camera for video, low end speakers, and a microphone. It will have Wifi, maybe one USB port, a built in battery, half a Gigabyte of RAM, a 4-Gigabyte solid state hard drive. Data input is primarily through an iPhone-like touch screen keyboard. It runs on linux and Firefox. It would be great to have it be built entirely on open source hardware, but including Skype for VOIP and video calls may be a nice touch, too.

Overall it’s pretty sound though a few things jump out at me.

• Battery LifeThe screen for something like this looks like it will be a decent size (9-12″), meaning it will need a decent battery. Getting a slim and light enough battery with enough battery life to allow for WiFi browsing (since that’s what the device is dedicated to) could be difficult. Your going to need more than an iPhone battery. I suspect under 4hrs will never go. 6-8 for any real adoption.
• Resolution – If it doesn’t hit 1024×768 it’s going to have a tough time being popular. That’s pretty much the standard most sites are made for.
• RAM – 512MB should be enough for a product that’s just a web browser / VoIP terminal, but if it left the ability to add up to 1GB (even if no easy access door) it may do better in terms of getting adoption.
• PC Card – The major criticism of the MacBook Air was the lack of a PC Card slot for a wireless card. You know that will be a case here. This could be tricky.
• Durability – Gadgets tend to last 2-4 years. We’ll go with 2 for the sake of this discussion. The front is a giant screen. It’s shaped like a book. Many people will want to carry it ina backpack or other bag containing other items. The iPhone is somewhat unique for using glass rather than plastic. The iPhone is tough as nails. If this screen gets scratched easily, it’s going to become a crappy experience overnight. If glass isn’t an affordable option, perhaps a low cost alternative is to make a cheap and easily replaceable plastic cover. So if it gets scratched up beyond the users threshold they can order another cover and just replace it.
• Software – I agree with the slim idea, but this runs into the same issues as Apple had with the iPhone. Web Apps don’t always cut it. The ability to hit the OS should be there (at your own risk) with an easy way to restore your device to factory condition (perhaps by connecting to the desktop and running some application). I know I’d like an SSH client (openSSH will be fine). Pidgin perhaps? Skype would be cool too. Easily hackable would be a major plus. Especially considering the nature of the early adopters.
• Stand – A stand with a built in USB hub and charger would be a very good accessory (keep cost of actual tablet low). It could be designed like a monitor so when you put the tablet in place, you can have a keyboard in front of it… and use it as a terminal with mouse/keyboard. Or just use the touch screen by tilting it back. All while it charges.

The ability to adopt some or all of these ideas needs to take into account price. But these are what I think will likely gauge it’s success or failure, assuming it reaches the market. The benchmark is the iPhone, love it or hate it. Being as user friendly, flexible, and durable is important. Taking advantage of the form factor, and a reasonable price is what will set it apart.

What will we call it? The best name I can think of is the Firefox Tablet, but that will take a round of discussions with Mozilla.

With the modifications to Firefox, that’s not likely. Get ready for IceWeasel Tablet.

Now that I gave my $0.02, I’d be curious to know what others think of the idea.