crosswinds
New Member
I've recently come under attack about offering no limit hosting. I've been lumped in with all the scum that oversell a $75 VPS after investing over $150k in an infrastructure to back up what it is I am doing.
Dan. In particular you have placed me in this group and I want to clarify. You have no idea how much I have been screwed over or sacrificed in an attempt to do what I have accomplished.
Read what I type BEFORE you respond. I am not here for a quick buck. I want to prove that no limits is possible, feasible and profitable. Once I have proof, I want to share - but it requires resources.
So here is HOW I have done it, and with what.
#1 - I do not use an off the shelf control panel. It is not possible to do it with any control panel that is tied to a single system
#2 - I don't use a single server or VPS. The initial system for CWahi (my free service) has 20 dedicated servers, over 10TB of disk in redundant arrays (both the disk chassis and their servers are redundant). Each system has a single function. Total hardware cost was around $75k with the last $75k being the 1 man year of figuring everything out and tying it together.
#3 - Figure out an expansion strategy for growth of b/w and disk. Both are easy given #2. More disk? Plop disk in the tray. They full? Buy more servers. Map them into the network. (zfs and freebsd for the win)
#4 - Map the users into server groups and design software to serve and do this mapping.
I have hand coded all the infrastructure to bind open source applications, servers and filesystems together. Admittedly, the members section sucks and needs to go more graphical, but it works, is scalable and fast.
No limit on disk space or file size - this is easy. Except for user created files or some open source projects, most large files are copyrighted. That is the one thing not allowed on our systems, so why limit them? Also removes one monitoring headache and introduces the need to watch for abuse. Old scripts from 1997 modified for the new layouts to the rescue.
Site getting big? No problem (right now this is not even an issue. The smallest server slice group is 1TB and the biggest site is only 18GB). But what if it goes to 1TB? Add more disk to the trays for the server group at 66% utilization, add the partition (zfs ftw). Done. Tray full? Buy new tray and server, image it, zfs the tray, add to network. Migrate some sites (live replication) and update their server group mapping.
Now the biggie - transfer.
No limit transfer is the game of making sure 95% of the people are not affected at *all*, while nailing the 5% that abuse/pillage your system (adult content is the biggest b/w consumer and against the TOS).
The total amount of b/w and transfer that a site can consume is not limited. This works FOR me. Fast sites to all normal end users (that 95%) is my goal - they see my text link bar at the top (in place of banners/popups) and they want to host with me. So I *want* to keep them (the normal people) happy.
The 5% that are abusers - this is the key issue.
So, using data from my pro hosting (when I had quotas and transfer caps), I find out what types and sizes of files are used in the construction of sites and most legal downloadable content.
I then write a system that applies a RATE limit to a visitor that goes out of this norm - and 99% of the time it is an abuser. No limits on the number of connections, just the max speed at which very large files of specified mime types can go out with.
95% of the people never see any limit. The 5% do, it sets off alerts, they get frustrated and leave, I find a site violating TOS.
I use a lot of B/W from my Colo and they have much more to give me. My prices are per Mbps is peanuts now.
The last piece is having sustainable revenue. And the trick is to try to not put it on the users pages (and I avoid the post 2 host method).
Ads are on the main site.
Ads are on the error pages.
I am also helping my ad providers monetize my biggest traffic base - China. Considered to be the hardest to monetize.
I average $6ish CPM - only about $20/day atm but my costs are approx $5-7/day (fluctuations of course). I haven't even actually paid for more b/w yet because I haven't even come within 50% of my commit level (bandwidth I pay for whether I use it or not).
So, a well thought out, custom infrastructure that is flexible, expandable, reliable and mostly, inexpensive to upgrade, plus low bandwidth costs and high advertising income (relatively speaking).
Even if i make no more changes, I can continue to grow and expand as I am. But since I am making money, I am developping and giving more.
I even removed the limits off my pro so that my paying customers do not have to pay more or ever worry about limits. I had approximate 10% near my old limits, but with the economic downturn I don't wont them to have to worry about over charges or outtages. My signup rate in the past 6 months has gone down with the economy, even with no limits. Greed is not the point.
I fight the overselling scum - I removed them if I found them abusing my systems and taking advantage of people.
So please, in a civil manner, question me. I love to be challenged and force to think out my planning.
I have, in a previous thread, expressed that I want to extend my knowledge and cost savings to Free Web Site Hosters (you guys) by GIVING you FREE VPS or Dedicated servers and leveraging my economy of scale to help you lower your costs, increase your reliability and provide better services to your customers.
With the attacks I really gotta wonder if it's worth dealing with the interpersonal hassles which often overcome all financial reward for both parties.
I am here to help.
Sorry it was so long but if I went into *gorey* detail, it'd be huge.
I can see why I am lumped into the overselling barrel, but you guys cannot see why I am not. I hate what they do as well - I've been screwed over and it sucks, and I won't do it to anyone else.
Dan. In particular you have placed me in this group and I want to clarify. You have no idea how much I have been screwed over or sacrificed in an attempt to do what I have accomplished.
Read what I type BEFORE you respond. I am not here for a quick buck. I want to prove that no limits is possible, feasible and profitable. Once I have proof, I want to share - but it requires resources.
So here is HOW I have done it, and with what.
#1 - I do not use an off the shelf control panel. It is not possible to do it with any control panel that is tied to a single system
#2 - I don't use a single server or VPS. The initial system for CWahi (my free service) has 20 dedicated servers, over 10TB of disk in redundant arrays (both the disk chassis and their servers are redundant). Each system has a single function. Total hardware cost was around $75k with the last $75k being the 1 man year of figuring everything out and tying it together.
#3 - Figure out an expansion strategy for growth of b/w and disk. Both are easy given #2. More disk? Plop disk in the tray. They full? Buy more servers. Map them into the network. (zfs and freebsd for the win)
#4 - Map the users into server groups and design software to serve and do this mapping.
I have hand coded all the infrastructure to bind open source applications, servers and filesystems together. Admittedly, the members section sucks and needs to go more graphical, but it works, is scalable and fast.
No limit on disk space or file size - this is easy. Except for user created files or some open source projects, most large files are copyrighted. That is the one thing not allowed on our systems, so why limit them? Also removes one monitoring headache and introduces the need to watch for abuse. Old scripts from 1997 modified for the new layouts to the rescue.
Site getting big? No problem (right now this is not even an issue. The smallest server slice group is 1TB and the biggest site is only 18GB). But what if it goes to 1TB? Add more disk to the trays for the server group at 66% utilization, add the partition (zfs ftw). Done. Tray full? Buy new tray and server, image it, zfs the tray, add to network. Migrate some sites (live replication) and update their server group mapping.
Now the biggie - transfer.
No limit transfer is the game of making sure 95% of the people are not affected at *all*, while nailing the 5% that abuse/pillage your system (adult content is the biggest b/w consumer and against the TOS).
The total amount of b/w and transfer that a site can consume is not limited. This works FOR me. Fast sites to all normal end users (that 95%) is my goal - they see my text link bar at the top (in place of banners/popups) and they want to host with me. So I *want* to keep them (the normal people) happy.
The 5% that are abusers - this is the key issue.
So, using data from my pro hosting (when I had quotas and transfer caps), I find out what types and sizes of files are used in the construction of sites and most legal downloadable content.
I then write a system that applies a RATE limit to a visitor that goes out of this norm - and 99% of the time it is an abuser. No limits on the number of connections, just the max speed at which very large files of specified mime types can go out with.
95% of the people never see any limit. The 5% do, it sets off alerts, they get frustrated and leave, I find a site violating TOS.
I use a lot of B/W from my Colo and they have much more to give me. My prices are per Mbps is peanuts now.
The last piece is having sustainable revenue. And the trick is to try to not put it on the users pages (and I avoid the post 2 host method).
Ads are on the main site.
Ads are on the error pages.
I am also helping my ad providers monetize my biggest traffic base - China. Considered to be the hardest to monetize.
I average $6ish CPM - only about $20/day atm but my costs are approx $5-7/day (fluctuations of course). I haven't even actually paid for more b/w yet because I haven't even come within 50% of my commit level (bandwidth I pay for whether I use it or not).
So, a well thought out, custom infrastructure that is flexible, expandable, reliable and mostly, inexpensive to upgrade, plus low bandwidth costs and high advertising income (relatively speaking).
Even if i make no more changes, I can continue to grow and expand as I am. But since I am making money, I am developping and giving more.
I even removed the limits off my pro so that my paying customers do not have to pay more or ever worry about limits. I had approximate 10% near my old limits, but with the economic downturn I don't wont them to have to worry about over charges or outtages. My signup rate in the past 6 months has gone down with the economy, even with no limits. Greed is not the point.
I fight the overselling scum - I removed them if I found them abusing my systems and taking advantage of people.
So please, in a civil manner, question me. I love to be challenged and force to think out my planning.
I have, in a previous thread, expressed that I want to extend my knowledge and cost savings to Free Web Site Hosters (you guys) by GIVING you FREE VPS or Dedicated servers and leveraging my economy of scale to help you lower your costs, increase your reliability and provide better services to your customers.
With the attacks I really gotta wonder if it's worth dealing with the interpersonal hassles which often overcome all financial reward for both parties.
I am here to help.
Sorry it was so long but if I went into *gorey* detail, it'd be huge.
I can see why I am lumped into the overselling barrel, but you guys cannot see why I am not. I hate what they do as well - I've been screwed over and it sucks, and I won't do it to anyone else.
Last edited:
