d8888888b    888 .d8888b.888888888888888888888Y88b   d88P  .d8888b. 8888b
	       d888888888b   888d88P  Y88b   888    888        Y88b d88P  d88P  Y88b88888b
	      d88P88888888b  888Y88b.        888    888         Y88o88P   888    888888?88b
	     d88P 888888Y88b 888 "Y888b.     888    8888888      Y888P    888       888 ?88b
	    d88P  888888 Y88b888    "Y88b.   888    888           888     888       888  ?8bb
	   d88P   888888  Y88888      "888   888    888           888     888    888888   ?8bb
	  d8888888888888   Y8888Y88b  d88P   888    888           888  d8bY88b  d88P8888888888b
	 d88P     888888    Y888 "Y8888P"    888    8888888888    888  Y8P "Y8888P" 888     ?88b
About Me

The Condensed History of iaNett

Page 3: A startup is born

After being crammed into the corner of a spare apartment bedroom for over a year, our new office space at WSi almost seemed luxurious. The truth was, in fact, actually quite the opposite. At WSi, business was booming: the company was expanding rapidly, and new people were being hired nearly every day, so office space was scarce. In the end we were located inconspicuously in a small corner in the back of one of the offices, and this suited us just fine: we had no desire to be caught up in the everyday hustle of the office, instead prefering the relative peace of our secluded alcove.

We relocated the equipment from Shone's apartment to our new digs and began to settle in. Although we had been offered space in WSi's data centre in Harbour Centre (the main carrier hotel and interconnection point for Vancouver), we instead wanted to be near our equipment while we worked. The building had recently been connected by Group Telecom as part of their fiber rollout and we were given a 10mbps connection (which was quite significant at the time) and IPs to work with, enough for development. The plan was to later relocate the entire project to Harbour Centre where we would have access to substantially greater bandwidth.

Always content to make do with what we could find laying around, we managed to scrounge up a 44U rack, a large UPS, networking equipment, spare computers and various other pieces of hardware from different departments. So insatiable was our appetite for more equipment and our desire to save money that we became well known for repurposing equipment that was laying unused around the office.

However, we didn't always manage to achieve our goals; on one occasion, despite our objections, three 1U IBM Netfinity servers (each with two 600mhz processors and two 9.1GB SCSI hard drives) were purchased for us, and to make matters worse, when they were delivered we discovered that they were missing half of the RAM that they were supposed to include. Since RAM was so vital to the search engine, we struggled to figure out what had happened to it and to get it replaced, but to no avail. Our belief was that someone who had been involved in the purchasing process had redirected the memory expansion to another project within the company, but we were unable to prove this. Ultimately we did find uses for those three machines: they were nice and fast but we were very disappointed that the amount of money that those three machines cost would have bought us literally dozens of low cost, purpose built search nodes - diskless, fast CPU and lots of RAM - what we really wanted.

We put our frustrations aside and focused on development. We began by hiring a friend, Allan Scott, whom I had worked with before on several projects while we were in high school. Together, we debated over designs and implementation details, and gradually laid out the blueprints for a high performance distributed crawler, index builder and search software. We hired on two additional developers as well to help with the development process, and now that we finally had a team we forged ahead at full speed.

By June 2000, the new crawler was functional, and we had indexed our first million documents. The whole time, we were continually scrounging spare hardware from all over the office and by this time we had managed to collect a number of surplus computers which we set up as a cluster to run our crawlers, indexers and search nodes. Over the following months, we continued to improve and tweak the software - all the time being aware of the subtle changes beginning to affect our parent company. In many ways, we were privileged to be at arm's length from WSi: although we physically co-existed within their office and witnessed several major events within that company, we were in fact our own entity and thus largely unaffected by what was happening.

Home - Geocaching - PHP Scripts - Articles - Music - Programming - Hot Sauce Review
About This Site - Site Search - About Me - Link Directory - Contact Me

This site was constructed entirely by hand { more info }
Modified Tuesday, November 4th 2014 UTC

(C) Copyright 2000-2019 Marty Anstey ~~ I didn't rip you off, so don't rip me off.
server: ca-to-1