This is an interesting perspective from an actual service provider, but I still think you a missing a few points.
1) Moore's Law
Interesting that you bring up Moore's law since it applies to transistor density not network bandwidth. Note that internet bandwidth has NOT increased at the same rate as processor speeds. Client side hardware will get cheaper far faster than network bandwidth. This seems to favor client side PVRs.
2) Internet Access
While I am not disputing your particular business model, most people (in the US, much less the world) do not have access to high speed internet access (much less to a MAN). As such they will not be able to download good quality video files on demand, whereas at least half of the US is already recieving cable.
3) Infrastructure (again)
I am very impressed that you can serve 8000 users with 40 servers. I doubt however this will scale linearly. How will you do with 800,000 (something like Tivos current subscriber base). What about 8,000,000? As the complexity of the system grows, so does it's cost and its chance for failure.
4) HD Failure
I don't disagree, but what is the actual HD failure rate for Tivos? What is the mean lifespan of a Tivo harddrive? I have heard very few complaints so far.
To sum up, I agree that server side PVR/Video On Demand is the best solution in the long run. Unfortunately every business has to make it through the short run first. There have been several attempts at it in the US, but none has had the success that Tivo has. Time will tell if your business model can outpeform Tivo's in the short run. I suspect that Tivo will capture the client side PVR market, then use its advantage to migrate its users to server side PVRs.
I once felt the same way, but after some thought I think the client side PVR model is actually a good one.
1) Broadcast Transmission
Client PVRs rely on broadcast transmission to "download" shows. This means that hundreds of users can be recieving the same show on a local loop and it will cost no additional bandwidth. For server side PVRs to work, they would have to have enough bandwidth at each server to handle all attached users. This would be costly.
2) Infrastructure
To setup server side PVRs you would need to keep adding servers as your subscriber base increases. Not to mention that not every cable user has internet access on their line. This may cost even cost more than just buying a client side Tivo for each user. Tivo is losing money as is. There is no way they could have got the capital to support their user base if they had to maintain all the servers that would be required for server side PVRs.
3) Reliability
While you are right that hard drive failures would piss off consumers, hard drive failures are relatively rare. I suggest that server side PVRs would have even MORE problems than client side. Look at your average web site, which seems to crash everytime its linked to some certain web logs. And if server PVR crashes, then possibly hundreds of users would be pissed off at once, instead of just one.
Server side PVRs have many advantages in the long run, but the short run costs keep them prohibitive for the moment. A good solution may be to combine the advantages of client and server side PVRs. For instance keep the same Tivo functionality, but add the ability to (slowly) download a requested show from a server somewhere. You request it on Monday and get it Tuesday night or something. The new ReplayTV P2P show sharing might be a good model for this.
A lot of the problems discussed here involve a proxy that analyzes packets and does aim correction etc. Why is it so hard to to encrypt packets with SSL or somesuch? If my browser can do it why can't the average game? What exactly is the performance penalty for packet encryption?
tivo has downloaded a car commercial and most recently a sheryl crowe video/advertisement. the only difference between this and uk stuff is that uk is actually recording a broadcast show, whereas the US ones were downloaded through the phoneline during the nightly update. it shows up on the main menu as another option, but does not clutter up your now playing menu. also, the uk discussion group states that tivo will not override any existing show recordings to record these specials. i presume it will override your thumbs up preferences though.
Slashdot Mirror
1) Moore's Law
Interesting that you bring up Moore's law since it applies to transistor density not network bandwidth. Note that internet bandwidth has NOT increased at the same rate as processor speeds. Client side hardware will get cheaper far faster than network bandwidth. This seems to favor client side PVRs.
2) Internet Access
While I am not disputing your particular business model, most people (in the US, much less the world) do not have access to high speed internet access (much less to a MAN). As such they will not be able to download good quality video files on demand, whereas at least half of the US is already recieving cable.
3) Infrastructure (again)
I am very impressed that you can serve 8000 users with 40 servers. I doubt however this will scale linearly. How will you do with 800,000 (something like Tivos current subscriber base). What about 8,000,000? As the complexity of the system grows, so does it's cost and its chance for failure.
4) HD Failure
I don't disagree, but what is the actual HD failure rate for Tivos? What is the mean lifespan of a Tivo harddrive? I have heard very few complaints so far.
To sum up, I agree that server side PVR/Video On Demand is the best solution in the long run. Unfortunately every business has to make it through the short run first. There have been several attempts at it in the US, but none has had the success that Tivo has. Time will tell if your business model can outpeform Tivo's in the short run. I suspect that Tivo will capture the client side PVR market, then use its advantage to migrate its users to server side PVRs.
I once felt the same way, but after some thought I think the client side PVR model is actually a good one.
1) Broadcast Transmission
Client PVRs rely on broadcast transmission to "download" shows. This means that hundreds of users can be recieving the same show on a local loop and it will cost no additional bandwidth. For server side PVRs to work, they would have to have enough bandwidth at each server to handle all attached users. This would be costly.
2) Infrastructure
To setup server side PVRs you would need to keep adding servers as your subscriber base increases. Not to mention that not every cable user has internet access on their line. This may cost even cost more than just buying a client side Tivo for each user. Tivo is losing money as is. There is no way they could have got the capital to support their user base if they had to maintain all the servers that would be required for server side PVRs.
3) Reliability
While you are right that hard drive failures would piss off consumers, hard drive failures are relatively rare. I suggest that server side PVRs would have even MORE problems than client side. Look at your average web site, which seems to crash everytime its linked to some certain web logs. And if server PVR crashes, then possibly hundreds of users would be pissed off at once, instead of just one.
Server side PVRs have many advantages in the long run, but the short run costs keep them prohibitive for the moment. A good solution may be to combine the advantages of client and server side PVRs. For instance keep the same Tivo functionality, but add the ability to (slowly) download a requested show from a server somewhere. You request it on Monday and get it Tuesday night or something. The new ReplayTV P2P show sharing might be a good model for this.
sigless
A lot of the problems discussed here involve a proxy that analyzes packets and does aim correction etc. Why is it so hard to to encrypt packets with SSL or somesuch? If my browser can do it why can't the average game? What exactly is the performance penalty for packet encryption?
read his website. http://www.pdrap.org/
tivo has downloaded a car commercial and most recently a sheryl crowe video/advertisement. the only difference between this and uk stuff is that uk is actually recording a broadcast show, whereas the US ones were downloaded through the phoneline during the nightly update. it shows up on the main menu as another option, but does not clutter up your now playing menu. also, the uk discussion group states that tivo will not override any existing show recordings to record these specials. i presume it will override your thumbs up preferences though.
Excellent engrish for "More Than Meets The Eyes!" found on takara site.
Do all of magazines sell out?