Hi there,
I'm attempting to start a company (in Hawaii) and plan to feature several 1-hour animations on my website and expect (hopefully) to have 80k users downloading 2 animations per year. If I do the math, that means:

(700 GB animation x 80k users x 2 downloads) / 12 months = 18.7 TB monthly transfer

I have a few questions (if you don't mind):

1. What kind of web hosting service would actually be able to fulfill such bandwidth usage (at very reasonable rates -- I've only budgeted $1,500 per year for hosting)?
2. Can you provide services you would recommend?
3. I was assuming that streaming these animations would not be an option and that downloadable content is best -- would you agree?

So far, I've only been able to find hosts like DreamHost and ValueWeb that seem reliable and customer-friendly -- plus only charge about $0.02 / GB transfer. However, they only seem to provide up to 3.5 TB of monthly data transfer.

Any help/insight would be much much appreciated! Thank you.

For that level of transfer you may be better with an unmetered connection (i.e. you take xMbps bandwidth instead of a data transfer limit).

Also I reckon you'd easily need a dedicated server for this, so your budget is probably a little low. With a dedicated server you should find providers offering both methods of bandwidth/transfer limiting - i.e. if they only advertise the 'transfer limit' approach on their website, you should perhaps contact them for more details.

What you want is xMbps bandwidth, but the ability to burst up to some higher capacity (i.e. to handle traffic spikes).

I calculate that your usage would require (at least) a 60Mbps connection. This would assume that everyone queues up in an orderly fashion to download at the most suitable time (and people download around the clock). In reality you'd want a bit more than 60Mbps to handle this, and then the ability to burst to say 100Mbps for traffic spikes. I guess you'd release a new animation and then have a rush of downloads for it?

This raises a question... how will you be protecting the content (I'm assuming you're selling it?) to avoid/limit distribution via torrents etc.? If you're not selling it, this may well be a much more cost-effective way for you to distribute it?

Other than that, there may be some benefit in taking a smaller hosting plan initially (e.g. a VPS) and scaling up to something much bigger when the time's right. Hopefully this would work out much better for your budgeting since I presume/hope that as traffic grows your income would as well?

A VPS would be an ideal starting point since (depending on your provider) you can easily migrate to a dedicated server when needed, and from there it will be easy to scale up the bandwidth etc.

The potential difficulty comes when you require an upgrade to your dedicated server. This is difficult to do without downtime since you either have to move between servers (from an old one, to new one) or at least have downtime whilst components are swapped in your existing server.

Of course, the other issue in all of this is server management - are you likely to be able to handle that on your own, or would you require some/all of it doing for you? It's not just about the management itself (e.g. config/patching etc.) but also pro-active monitoring to prevent and respond to any potential problems.

With regards to streaming, I think this is your choice. Given your budget, I would probably say that downloadable content is better (since it will be cheaper); streaming will require much greater bandwidth availability and more diverse connectivity (i.e. to ensure a reliable connection to all visitors). To do streaming "properly" you also want to use a distributed service of the kind offerred by Akamai et al. but again, not an option given your budget.

I thinnk that webintellects, theplanet and a few more have unmetered server offers. But even with those you need to increase your budget to about $2000

For what you are asking and if you were to host it under a shared environment even if the host does offer that amount you are asking, I think you will have some porblems even uploading those files on a shared hosting account

Yep, just like damien_ls and aeromit suggested, you nid a dedi to do the job

Yes i agree with the same. He will definitely require a dedicated server. However, if he is actually looking for high-quality bandwidth. Then, he requires to think on his budget.

Thanks, guys! This has been really helpful. While I know some tidbits about IT, I'm obviously no expert.

As far as protecting content, I was planning to use Windows Media Player's DRM capabilities and have people download the animations since that's the cheapest alternative. However, I'm not truly sold on DRM: 1) It turns people off (often in a big way), and 2) There are always ways around it. So, what I'm thinking now is featuring the animations in streaming FLV format, since it's less obtrusive and is truly on-demand. Of course, this means my budget could balloon to $8k-$9k (http://www.flvhosting.com/index.php?sc=16), but maybe it's for the best.

If you happen to know anyone that beats FlvHosting's service, please let me know.

Added Note: I just found a hosting company called "InMotion" that appears to support streaming Flash and 9,000GB transfer for a ridiculously low monthly payment of $18.95/month. Any thoughts on this?

Have a great weekend!

Well, honestly speaking. You should be looking for high-quality space and bandwidth. Cheap is never a good solution in the long run.

Originally Posted by BariCarroll

Thanks, guys! This has been really helpful. While I know some tidbits about IT, I'm obviously no expert.

As far as protecting content, I was planning to use Windows Media Player's DRM capabilities and have people download the animations since that's the cheapest alternative. However, I'm not truly sold on DRM: 1) It turns people off (often in a big way), and 2) There are always ways around it. So, what I'm thinking now is featuring the animations in streaming FLV format, since it's less obtrusive and is truly on-demand. Of course, this means my budget could balloon to $8k-$9k (http://www.flvhosting.com/index.php?sc=16), but maybe it's for the best.

If you happen to know anyone that beats FlvHosting's service, please let me know.

Added Note: I just found a hosting company called "InMotion" that appears to support streaming Flash and 9,000GB transfer for a ridiculously low monthly payment of $18.95/month. Any thoughts on this?

Have a great weekend!

You're note probably isn't a good idea. What places like this do is either scam you directly or say you're taking to much server resources (as in processing, etc) once you get to a certain point so you can never use as much bandwidth as they say.

To save yourself some money in terms of server management, I recently saw this site & the reviews seem really good:

http://platinumservermanagement.com

Basically for $29/month they'll take care of, it seems, all aspects of management of your server (provided it's the OS/Control Panel that they specialize in).

I've never used the company, but when I move to a dedicated server I will definitely try it out.

Using a server management company, you can pickup a self-managed dedicated server (often a lot cheaper than managed servers), and pay a very small fee to have it managed for you. Maybe it'll help with the budget.

that is 20tb per year... and pure downloading only shouldnt be issue for those terabyte hosting provider like hostgator and bluehost etc.

Simply as they are giving 3tb per month.

Oversellers will suspend you the minute you get your site up. For 5.95$ you can't expect to get as much as a dedicated.

Well your budget is kind of small for what you need.

What you could do is go for a streaming server that streams data for mediaplayer. This way you won't get the high loads of website streaming and you can concentrate on getting as much unmetered bandwidth as possible.

I would also recommend you start off with 10-20 mbps and upgrade when needed.

The good news is that I think I can save you some money. The bad news is that, in doing so, I'm going to tell you something that you probably don't want to hear.

You don't actually need the 18.7 TB of data transfer right away. You're not going to go from 0 to 80k users in 5.2 seconds or anything like that...in fact, you will probably never reach that goal in all likelihood. I'm sorry if that isn't what you want to hear, but you need to find 80k users willing to sit and wait for a 1-hour animation to download, even to the point where it can be streamed and played without being choppy. That probably isn't going to happen any time soon. I know it's your project, and you want to aim high, but right now you seem to have rose-colored glasses on and it may be time to take 'em off.

My suggestion to you would be to take your traffic goal and cut it to about 10% (in other words, 8K users...and even then, you're pushing it) for the first year, and find a company that will grow from there.

The other suggestion I have for you is to budget for a colocated box and pay for the bandwidth. It will probably be cheaper in the long run.

Originally Posted by ADAM Web Design

You don't actually need the 18.7 TB of data transfer right away. You're not going to go from 0 to 80k users in 5.2 seconds or anything like that...

<snip>

My suggestion to you would be to take your traffic goal and cut it to about 10% (in other words, 8K users...and even then, you're pushing it) for the first year, and find a company that will grow from there.

Yep. I completely agree that it'd be a waste of money to take capacity for an ambitious target like this. However, if you have good reason to believe that you can reach your target it would be wise to ensure any option you take has an easy upgrade/migration path to get you onto the optimum solution for the big traffic you desire.

Originally Posted by ADAM Web Design

The other suggestion I have for you is to budget for a colocated box and pay for the bandwidth. It will probably be cheaper in the long run.

I see this as your optimum solution. It will be the cheapest when you get there, and may even be possible to start with. Otherwise, you want to take a look at how you will get from your starting point to a colo server with the minimum fuss and disruption etc.

With regards to streaming, I would suggest it's best avoided unless you're taking a service from a big provider with a proper distributed streaming service such as Akamai. This level of service doesn't come cheap, and you can expect to spend the kind of annual budgets you've mentioned within a month.

The cheaper way to do it is via download - although it is possible that streaming would work out a more attractive option for your customers/visitors and/or have other marketing/customer insight advantages?

As everyone else, I too would recommend you to start with a shared hosting. Checkout lunarpages. They offer 3.5 TB bandwidth.

Once you cross the limit, get a dedicated server with unmetered 100 GB port. That will be costlier, but you don't have any other option.

Originally Posted by damien_ls

I calculate that your usage would require (at least) a 60Mbps connection. This would assume that everyone queues up in an orderly fashion to download at the most suitable time (and people download around the clock). In reality you'd want a bit more than 60Mbps to handle this, and then the ability to burst to say 100Mbps for traffic spikes. I guess you'd release a new animation and then have a rush of downloads for it?

This is a rather strange assumption that during traffic spikes you need only 1.5 times more bandwidth. In my experience traffic during spikes can increase a thousand times. And the spikes depend not on adding new materials to the website, but on unpredictable behaviour of blogosphere, avalanches of blog entries, referring to the materials from the site.

Why not use a vidoe content publishing system such as veoh.com, which allow you to charge pay per view. Granted you'd have to share your revenue with them, but you can test out the market for very low initial investment (just your site without vidoe streaming which you host anywhere for $5 - $10 per month), it will help you figure out the demand for your animation videos, later you can create your own video streaming site to avoid sharing revenue.