Continuing the series - it requires a fair bit of hardware to churn through the traffic we do deliver.

We started off with the traditional one server setup - providing web, sql, DNS, mail, and what not all in one box. That worked for a while, but pretty soon we were running at capacity. Our solution was relatively simple - push our non-login based services (such as PageRank Checker and Button Maker) onto a secondary server. It was a simple split, and we wouldn’t have to worry about SQL replication problems. Our actual blogs were on another server, and Blog Top Sites had its own server.

As we grew, it become evident that this solution wouldn’t pan out. By now Enthropia was growing at a rapid pace, a decision was made to invest into a reliable infrastructure. I posted about things to worry about when having your own farm. We ended up with 12 dual dualcore opteron 265s. One of them was the iBegin database server, housing 16 GB ram and 8×72 GB hard-drives.

As previously mentioned, our previous setup of half a dozen servers consisted of solitary servers. Each had to deal with DNS, email, web requests, and SQL lookups individually. This meant a lot of wasted overhead running processes that weren’t required. In our new system, one of our servers functions for both DNS and email. As we grow, adding a secondary server to handle DNS is elementary.

Each server now individually runs apache+SQL. While it may seem like a bizarre decision (we could have gone for a master-slave SQL setup with the other servers acting as web-servers), it makes things simpler for us. Because everything is internal to our own network, whenever we need to do cross-server requests (such as MapStats making sure a user is logged in), it is no problem.

We then have daily weekly and monthly backups to a backup server provided by our host (it can store upto 9TB). We also do monthly DVD backups, as a just-in-case emergency deal.

All-in-all, we utilize five servers for Bloggy Network. We are running at roughly 30% capacity, giving us a lot of room to grow.

I’ve been meaning to write this post for a while … but my friend Mike Bogo summed it up better than I can.

He is spot on - more than anything else, people seem content on regurgitating what someone has already said. I desperately want to find more meaty blogs - ones that provide actual analysis. It is okay to be a few days late provided you bring something to the party!

In the last 3 months I’ve added one new blog to my feed: DomainTools Blog. These guys have interesting numbers that no one else have, and they provide some insightful look at it.

I’ve reached the point where I’m deleting about 9 out of 10 trackbacks. Their entire content is roughly this: “Ahmed has talked about Point X, click here to read about it”

More original content please

Continuing on from Bloggy Network: the series the first site I want to talk about is Blog Flux.

As I had explained in my previous post, Blog Flux arose from the ashes of EatonWeb. What it did differently compared to all other directories was that it eschewed the traditional hierarchal structure, instead opting for a free-flow category/tag structure. Choose upto 10 categories - it is not my position to tell you what your blog is (or isn’t) about. Initially we had allowed an unlimited number of categories, but due to abuse, it was chopped down to 10.

My original plan with Blog Flux was always to be more than just a blog directory - it was to be a clearing house of anything and everything related to blogs. Having to bounce through multiple sites for simplistic services seemed a bit cumbersome. Bloggers were forced to go to X site for a button maker. And then Y site to perform pings. And then Z to see their statistics. The entire process was extremely cumbersome - remembering multiple logins was a headache.

Not everything went perfectly. We had to retire two subsections - Subscriptions and LinkLog. Subscriptions we had launched just around the time FeedBlitz. was launching. While I think our solution was quite lightweight and simplistic, FeedBlitz was honestly just the better solution. The other one that went down was LinkLog, similar to the original function of MyBlogLog. We extended the system, tracking AdSense clicks, and adding in a lot more functionality. Unfortunately, this was something that collapsed under its weight due to issues with scaling. Dealing with 100,000 clicks a day is easy. Dealing with 1 million clicks a day becomes much more complicated!

While we originally launched services in a rapid manner, what has really slowed us down now are the beforementioned scaling issues. With the size of the operations now, we need to test each system under heavy stress. For example, the Who’s Online system took about a day to write. But what wasn’t so easy was scaling the entire options - it generates over a million pageviews a day. Handling that load with speed requires more finesse than programming 101.

Overall though, Blog Flux has been a huge success. The site is simplistic, and offers pretty much whatever you want under one roof - from statistics to easy tools to blog themes, it requires one login to utilize them all. And users love it and are more than happy to link back - we have 10 active subdomains, and two defunct (Subscription + LinkLog). Every single subdomain that has PageRank has a PR of 7. Google reports 217,000 backlinks, and Yahoo clocks in at 2.9 million backlinks. The site itself generates 1.25 million pageviews a month, but we serve up over 20 million apache requests per day!

There is more coming, The integration of Commentful should be ready next week. We are working on a local blogging system that should launch in 4-6 weeks (powered by iBegin Source. Within the year I expect us to hit 2.5 million pageviews a month.

The title says it all. This post by Domain Tools goes into more details on how Karl Auerbach, a former ICANN board member, told Name Intelligence (who operate Domain Tools) that the cost to VeriSign is far lower than they claim. This goes right along with my point that due to scaling, the cost per domain goes down, not up.

The other great point they raise is domain tasting - the ability to register a domain for five days and get a full refund. If the DNS was taking such a beating (as VeriSign claims), why even allow it? That would make a noticeable dent on all the registrations going on.

($6.00-$0.14)/$0.14 = 4200% markup = 97.6% profit (old).
($6.42-$0.14)/$0.14 = 4500% markup = 97.8% profit (new).

Poor VeriSign, they aren’t happy with their ‘little’ margins.

Updated: Steve correctly pointed out a flaw in my wording. Furthermore, for people with the typical “Don’t use them” defense - if you buy a .com or .net, you have to use them. Read my previous post.

Bloggy Network (the resultant of a merger) is now roughly 15 months old. It does roughly 4.5 million pageviews a month now, and should crack 5 million pageviews next month (growth is in fact accelerating, not tapering off). All organic without a dime of investment money.

My next few posts are going to be about Bloggy Network and the (many) challenges we have faced. My first post (this one) is going to be about our history, and how we started.

Enthropia Inc has been around for a while (registered corporation for over four years now). As a web-dev company, we pretty much only develop projects for in-house usage. Occasionally we decide we no longer want to be part of a certain market, and jettison our investments (ie sites) in that niche.

I first came across Jacob Gower February 26, 2004, when we were selling one of our sites. He was quite happy with his purchase, flipping it for a very handsome profit a couple of months later.

We kept in touch, with Jacob buying most of the sites we sold (including one of the few sites I regret selling - Webmaster.org). I remember him telling me about having to sell Webmaster.org as he was committing himself to blogs. As we kept on conversing, he kept telling me how fantastic blogs were. Around this time we were still involved with Evo-Dev, and we decided on a joint venture: Blog Top Sites. We had a lot of experience in running topsites, and Jacob was involved in the blogosphere. The site was an instant success, signing up hundreds and hundreds of blogs in the first week. Today the site tracks over 4.5 million pageviews a day across over 40,000 blogs. The site generates over 30,000 pageviews a day now.

As things changed and aligned themselves in the right direction, we ended up purchasing EatonWeb, the oldest blog directory out there (side fact: Brigette Eaton was one of the first people ever fired for blogging, somewhere in 1999 I believe).

The site was popular, but it was horribly coded (PHP3 in fact). I also took a look on the blogosphere - it was extremely fragmented. We did what we do best - launch a new site, accumulating many of these fragmented services into one place. Only this time it wasn’t from scratch - we had Eatonweb to help push it forward Our original launch was nothing more than a blog directory, button maker, and pagerank checker.

Success was at hand. By eschewing the traditional hierarchal structure in web-directories and choosing an open multi-category/tag format, we were doing directories ‘different’. People liked having multiple services under one roof, and we expanded our offerings.

By now Jacob had designed his network ‘Bloggy Network’ (a name suggested by me). He had a few good blogs under his belt, including Forever Geek and Blogging Pro. We had 50% of the largest blog topsite (which spawned dozens of duplicates) and we had the best blog services site.

Enthropia has roughly 6 divisions. Each has its own manager. So when thinking about our future plan into blogs, there was only one logical person: Mr. Jacob.

And the pieces just fit.

Next up: Blog Flux, and why it clicks.

Ugh. I just read that VeriSign is increasing the base cost of .com domains 7%, from $6.00 to $6.42.

Some background: A while ago, ICANN and VeriSign sued each over. In the settlement (as VeriSign has tons of cash, and ICANN doesn’t), ICANN agreed that VeriSign could continue to be the exclusive provider of .com/.net for the next seven years. Furthermore, VeriSign was given the ability to increase the base cost (which at that point was $6.00) upto 7% every year. The word ‘upto’ being useless, as how many corporations decide to go for 5% extra profit instead of 7%?

Anyway, VeriSign, being the opportunistic monopolistic company they are, jumped at the first chance to make some more money.

This is a company that has an exclusive lock on .com/.net domains. For every year I renew this domain, they make $6.00. Now instead, they will be making an extra 42 cents. Going by TechCrunch’s math, that means $27 million extra per year. For doing nothing.

What is rather disheartening are the people spouting out comments over at TechCrunch without having a clue what is going on. The basic arguments:

• Domains used to be $100. Thats great - they used to be free. Stupid argument
• Don’t use them. I don’t know what people are smoking - if you buy a .com or .net, VeriSign will get your money. The low-cost registrars still pay that fee. And they had nothing to do with how ‘cheap’ VeriSign is (ie - they did not exist until the price went down, not vice-versa)
• Domainers. Love them or hate them, it is Google and Yahoo that give them the money on their PPC pages. $6 vs $6.42 - you think that is going to stop domainers?
• Infrastructure. This one really irks me. Do you remember the cost of hosting 10 years ago? 5 years ago? Newsflash: technology makes things cheaper and cheaper. VeriSign states 30 billion DNS queries a day - and yet OpenDNS does over a billion without much problems. Somehow .info operators are able to run a registry at $1-$2 per domain, and that is with much smaller volume. Scale up from 5 million domains to 65 million domains, and the cost per domain goes down, not up.

What is just disgusting is how people are actually defending a monopoly. VeriSign answers to no one (well, they are supposed to answer to ICANN, but just observe the Registerfly debacle and how incompetent they were at that).

If something costs me 50 cents, and I used to charge you $50, but now charge $5 - I’m not doing you a favor - I’m still ripping you off.

None of my sites have broken this mark, so it feels good. From my previous post on why Google is #1, that same site has been hit 100k times in one day:

Just saw that another local search engine has been released: YellowBot.

And so another chapter in my ‘reviewing and criticizing local search sites’.

I will first send our attention to this: Yellowbot.com - Social Networking and Local Search. Why don’t we go over their feature-list:

Local Search - An obvious one. Dozens of sites for that
Advanced Search Filters - I see no ‘advanced search’ option. I’m going to chalk this up as buzzword sensation
Tagging - Been there, done that with iBegin
Ratings and Reviews - The staple of all local sites
Recommendations - The only thing I saw was ‘recommended - yes or no’. Seems like simplified version of a review. Similar to our own ‘Favorites’
Business Details - Unless details are the basic name + address + category, I missed something.
Uploaded Pictures - Another staple
MyAccount Page - Again, another staple
Blog - I don’t see a way for users to blog, so I imagine this is the company blog. Old hat
Forum - Another staple (be it in the form of ‘Talk’ or ‘Community’ or even ‘Q&A’ - same thing)

Other things to nitpick:

• When adding a review, a nice inline popup appears, and then a nasty redirect to the login page. Why even bother with that guys?
• http://www.yellowbot.com/tags/food/El%20Paso%2C%20TX%2079922/page1.html is not a proper URL. You need to add url encoding
• On the map page of this listing, what’s with the ‘Good to Know’? (plus you wrote it as ‘el Paso’ heh)
• Your MapView is busted - no icons for above ‘100′ it seems. It would be prudent to use the ‘bounding’ function in GoogleMaps so it auto-zooms in as close as it can

In your WebProNews interview, you said people do not get the connection between offline and online. In what ways are you guys making that connection better?

Aaron from Find Buffalo has a dual-interview up, where he asks Paul Jahn and Matt McGee about local search et all.

Do notice how Paul mentions iBegin Source - rawr!

I was reading the latest Business 2.0, and in it was an interesting article on Various, the company that owns AdultFriendFinder (amongst many other *FriendFinder.com sites). It also had an article on LibraryThing, where people can pay to list books they own (the first 200 are for free).

An interesting point by the owner of Various, Andrew Conru, was about how people wanted to pay - in the case of their BigChurch.com site (religious dating), their registration jumped immediately after adding a subscription fee. To quote verbatim:

“On the Web there are not enough filters to sincerity”

I couldn’t agree with this quote anymore. As is evident on social-networks like Digg, the ‘mass’ effect is not necessarily a good thing - mob mentality caters to the lowest denominator, with fantastic news stories such as ‘AWESOME PIC’ the current du jour.

Money has always been a filter. Be it a private school that asks for a ‘donation’, be it a conference you have to pay $5000 to get into, or be it a forum that charges you access to its member areas.

The evolution of most medias has evolved from free -> advertising-supported -> subscription based.

It has happened to newspapers
It has happened to radio
It has happened with television
It has happened on the web (ie online newspapers or forums)

We already have a Digg-like site with a paid barrier of entry with Metafilter.

There are two fields that require a subscription/upfront-fee for admittance (and its coming):

1. Blogs
Yep, I said it. I remember when Greg Sterling mentioned the idea (sorry can’t find the link), I was ready (quite a few people were not). And why not? What is so different about blogs that a subscription model is so horrible? Heck you can argue Search Engine Watch already does it.

2. Local Reviews
I’m not advocating something closed off like Angie’s List. No - the entire point is to leave everything open, but - to contribute you must become a subscriber. It’s an odd situation isn’t it? Everyone can see everything for free, but if you want to write a review (which takes up your time), you gotta pay for it.

Lets step back for a second and think of the situation (where the initial catch-22 problem has been solved). You have a local review website that has 10,000 reviews in your area. Every single reviewer has been ‘qualified’ by having to pay an upfront free (lets say $15). In such a situation, people want to contribute. And knowing they will be contributing alongside serious people (qualified by the buy-in), the ensuring contributions should also be high quality.

There will always be shills. But a combination of upfront free + ‘invites’ could be a great way to go at it (if done properly).