Welcome to WebmasterWorld Guest from 184.108.40.206
Here is my goal. I want the highest quality, fastest loading time possible with the minimal amount of dropped video frames. (example, check out the spiderman trailer, I get no dropped frames when I watch that, but on You Tube all I get to see is about 4-6 video frames per second, yech!). I am perfectly fine with limiting the number of people who could be on my site at any time. I don't need nor expect to ever have 50 people looking at my site at a time. I assume if I can have 10 at a time max that would be fine and perhaps way more than will ever be looking at my site. If my server crashes, what recourse do I have. In a way I'm really just trying to have a glorified random access video player that can accommodate up to ten people at a time.
Can my above goals be done from home by buying my own server? Can I achieve the fastest possible download times if I limit the total running time of all of my films to an hour max, or is that considered a lot? Will the Electricity bill be huge? Am I saving money if I can actually pull this idea off versus finding a host? Is it a time passage issue, and if I keep the system long enough, it eventually becomes cheaper, how long for that threshold point arrive? What is the minimum amount of bandwidth I need to achieve my goal?
The videos would have some kind of website around them, but probably nothing more than 10 "pages" worth.
Is anybody here already doing this exact thing I am considering doing? Could I set it up so the first viewing is free but then additional viewings cost?
Thanks in advance.
over the years I have purchased expensive video equipment, used it for a decade, and then obsoleted it.
Servers become obsolete a whole lot faster than that.
Can I amortize the cost of a small server over several years
In the U.S., you can depreciate it over a small number of years, or you can (depending on your individual situation and other deductions) most likely write it off completely in the first year as a Section 179 Expense.
and keep it where I live?
Probably the worst thing you can do in your situation. Your bottleneck is bandwidth, not server horsepower. The least likely place you are going to get the needed speed and quality of bandwidth at a reasonable cost is in your home.
Can my above goals be done from home by buying my own server?
I'd guess that buying your own server will satisfy none of your goals. Assuming you are not going to have the server transcode the video on the fly, the server resources needed for this are minimal. The server is not the limiting factor.
I doubt very much you need a dedicated server. I think a VPS would be fine for this. Make sure you get one where you get a guaranteed percentage of CPU, and ask what CPU chip and speed they are using, and how many VPSs are on a computer.
Even if you did need a dedicated server, you are much better off leasing one from a host than buying your own. It will be professionally-managed, have backup power, and be located in a data center where suitable bandwidth will (hopefully...) be available. And you will have no obsolescence problem.
Can I achieve the fastest possible download times if I limit the total running time of all of my films to an hour max
The total running time of all your films is irrelevant. The number of simultaneous viewers, required bandwidth, and available bandwidth are what matters.
Will the Electricity bill be huge? Am I saving money if I can actually pull this idea off versus finding a host?
It will cost you significantly more to host this in your home than to use a host. Not because of the electricity bill, but because of the cost of bringing suitable bandwidth into your home.
What is the minimum amount of bandwidth I need to achieve my goal?
We can't answer that question. You can. Encoding rate x number of viewers.
But you don't just need "n" bandwidth. You need "n" bandwidth that can actually reach your users. You need to be hosted in a data center that has multiple large pipes to multiple backbone carriers, so that a reasonably-optimal path is available to most users. (And your host has to actually USE them... some hosts can be cagey about this.)
There are also services that will push your videos to "the edge" of the network using caching servers in multiple locations. For something like a high-quality Spiderman trailer, this is how they would do it. I don't know if this makes either practical or economic sense for you, though. Companies that do this include Akami, Digital Island, Mirror Image, Sandpiper Networks.
I don't understand the principles involved in streaming video (which is why I posted on this forum). What if I set up five or 10 digital harddrive machines that stored my videos on them, sort of like a video jukebox, and whenever someone wanted to see a video, they would actually be activating one of those machines via the internet. I suppose that idea is probably flawed and probably less flexible than other ways of doing it, its just that I like the idea of knowing exactly how my product is being streamed and I don't like the idea that I really won't know who, what. where or when is actually making my videos run.
It appears the biggest problem to doing it myself is the pipeline from home, but what are we talking here? What is the size of the pipeline that I need to have random access video of some short films?
I don't know if there is a reluctance to share the information I am asking for because perhaps it goes against the "industry" outsourcing, or perhaps it's because you already know it's not a wise way to go, but the way I look at it is if I can get the best information on doing it myself, then do it myself, I will then have a base of knowledge to work with which I currently don't have.
After my do it yourself adventure I probably will discover that hiring a company that specializes in what I am after is a much better way to go, but at least now I will have experience and an understanding of why that is and most probably I'll spot when I'm being giving poor information and poor service versus when I'm being given good information and good service.
This is for the most part a faceless industry, not being able to walk into the facility that is handling my account sends a big warning signal up to me and the only way to avoid it is to either find a company near me, or really understand every step of the process.
[edited by: ItsAlex at 7:28 pm (utc) on May 7, 2007]
What if I set up five or 10 digital harddrive machines that stored my videos on them, sort of like a video jukebox, and whenever someone wanted to see a video, they would actually be activating one of those machines via the internet.
ONE computer is WAY MORE than enough for this! Really! The server is going to be doing practically nothing to serve 10 video streams.
Again, you are the only one who is going to know what bandwidth you need. Experiment with codecs and bitrates and decide what is going to give you acceptable quality. Then multiply the bitrate by the number of simultaneous streams you want to be able to serve.
Running ANY kind of webserver - but in particular a video server - out of your home is a really bad idea, and you are going to find almost nobody here who is doing so.
I think you need to begin at the beginning and start learning the basics of web hosting and creating websites. Serving videos isn't much different than serving anything else - except that it will suck up much more bandwidth. ;)
I'd go along with the advice to first build a site and take things one at a time - you can use that site to learn how to incorporate a video, get an idea of how much bandwidth you're going to need, learn to set up a payment system, etc., etc. And if you want to know all the details of how your videos are being used, most hosts have ways of keeping track of the stats for you.
As you work on the site, read. Read a lot. The questions in your post would spread out over at least four of the forums on this site. As you read, you'll start to get an idea of what kinds of things you need to learn, and as you get some hands-on experience with a site and run into specific problems, you'll be able to ask specific questions that people can help with.
[edited by: Beagle at 11:28 pm (utc) on May 8, 2007]
And I'll try to pick things up on these forums as well.
You want to host them from your house - so you'll be essentially using a domestic, or domestic dressed up for business connection. This will have a quoted 'upstream' speed. For ADSL in the UK, that's something like 300kbps.
Ok, so how many people can view your streams at once is (roughly) 300/128 = 2. And that would be in ideal circumstances, and it would mean that your connection to the internet would be pretty impaired if you actually wanted to use it.
If you were with a hosting company, then you have the possiblity of a *lot* more bandwidth to play with.
Home hosting also means you'd be doing all of your own system maintenance, thinking about what to do in the event of power outages etc.
Network cards can handle a boatload of traffic, a lot of servers would have a hard time filling the volume of traffic the network card can handle.
So you've got a server that isn't doing much, sending it out over your network card that's basically twiddling it's thumbs. It's right where the network card touches the outbound connection that your problem starts. If you want to serve 100 connections, you're going to need big pipes to send all that data through. That makes 'big pipes' likely your primary concern. Home connections do not have big pipes; data centers do. That's why even a shared server will probably do OK, as long as you've got a big connection to the web.
I don't know much about videos online, but I believe I've seen specialized hosting companies that do this. Might be worth checking those folks out.
I wrote a thread about it a little while Back.