[DaJuice]

I'm not fully up on SDD and flash drive technology, but I don't think transfer rate scales with the number of chips used. I guess you could do striping to attain those speeds, except that would leave your proposed unit with only 16GB of storage. As for the kiosk idea, ehhh... very un-appealing to me personally from a consumer standpoint.

[eastmen]

Originally Posted by specwarGP2

LOL. Can you name at least 3 different experiences that you cannot recreate on an optical disk other than faster loads (since there are so many of them)?

Your arguments are starting to get a bit eccentric, to say the least.

Its up to the developers to use it to create new experiances. however with such fast transfer rates you can create true open worlds that aren't the same textures used over and over again creating bland worlds like gta4. Having large storage capacitys don't mean anything if you can't get that data to the machine fast enough and if we do see 4 gigs or even 8 gigs of ram with 54MB/s it will take you at least 74 seconds to fill up 4 gigs and 148 seconds for 8 gigs . Thats not counting seek times or the time to get a disc to actually spin up to 12x speeds.
With a 240MB/s transfer speed it would only take 16 seconds to stream in another 4 gigs of ram and 33 seconds to stream 8 gigs of ram. Seek times will also be low and you can constantly stream that data over. The only thing that will limit you is capicty and costs . But imagine having 50 gigs of unique textures that can be constantly streamed to system ram. The world you can create will be amazing .

An example is Call of duty. You call a air strike and depending on where the physics dictate damage done and buildings colapse in a multi player map you can have unique textures stream in and show diffrent outcomes and allow the players to continue using the map. Something of that scale would be limited to much less with only a 54mB/s transfer speed. Of course if you could get a 1GB a second transfer speed a game designed around that would be limited by 240MB/s transfer rate. Developers ambitions will grow as they are offered new technology

[aaronspink]

Originally Posted by eastmen

Dvd's don't even cost pennies. Dual layer dvds i believe are still in the quarter range. Regardless , with flash ram soultions you will gain speed , capaicty and seek time advantages .

yes but that's a double pressing requiring over 2x the machine time, single layer BR will be cheaper than DL-DVD if your volume is high enough to cover the master cost. Single layer DVD is in the <$.10 in with volume contracts.

[Unknown Soldier]

Originally Posted by eastmen

I dunno , here in Jersey we just got Fios and 20Mbps is cheap but your looking at almost double the cost for 50Mbps. I doubt we are going to see 100mbp at an affordable price before next gen consoles come out. Also when looking at 50GB games or even 25GB games those will not be fast downloads.

Ahh, but you seem to miss something. With that 20Mbps you should really be getting +4Mbps of true download speeds. That's more that enough. I've downloaded content off the net that are around 8-12GB and it takes 30min. for 8GB at true 4Mbps speeds. So, in my experience I believe that lovely 20Gbps you say you have is already enough for now.

Here in SA, the basic package consists of 384Kbps and 1GB download/upload(total traffic including International traffic i.e. UK/US) content. Now up to 2-3 years ago, they were charging around 50 Euros for this if not more. Things have gotten better now and for that package you now pay around +-25 euros. For the max package, you looking at 4Mbps(shaped) and up to 10GB. you looking at paying around +-70 Euros for this. Problem I have with this is that I feel that 10GB is not enough. First, the 4Mbps allows you to download content(games, mp3's) at a faster rate and you will reach your cap within a day or two if your not careful. Now, most of the people play games every night(4-8hours) and can use up to 400MB just playing games, times that by 30 and it comes to 12GB .. so a person who plays games every day will be capped before the month ends, remember since here they total to download and upload together to get the limit(which is something else I have a problem with). What if a person wants to browse or down a new game through Steam(usually around 4GB if not more). You see the problem now? Then there's shaped and unshaped packages. Most packages are shaped, meaning that the traffic internationally is worse than if you buy the unshaped package which is a lot more, well over 100 Euros(i'd guess about 120 Euros). So a person who wants to play WOW, and wants good pings 200ms(UK) or 260ms(US) needs to get and unshaped account.

And this is all because the government has/d a policy where it protected the telco which incidently the government has shares in. Now they decided about two years ago to get a Second National Operator(SNO) but again, the government has shares in it and well, things are only slightly better than 3-4 years ago but not really helping at getting the prices lower or helping at getting better, faster internet/bandwidth services.

US

[eastmen]

Originally Posted by aaronspink

yes but that's a double pressing requiring over 2x the machine time, single layer BR will be cheaper than DL-DVD if your volume is high enough to cover the master cost. Single layer DVD is in the <$.10 in with volume contracts.

Doesn't bluray require extra steps including that coat of anti scratch resistant stuff they put on it which would keep it more expensive than dvd ?

[Unknown Soldier]

Originally Posted by -tkf-

To get that speed it usually requires a heavy investment since Cobber only goes so far. And with every insane speed bump to the private users the backbone has to be buffed as well. In 10-15 years we may have "instant" access to movies, etc..., but it will still be limited to those lucky fews that live the right place with the right ISP.

I don't believe it'll take that long. I say 4-5 years. for the tech to become more accessible. Also remember that yes, the technology is expensive for the ISP in the beginning, but the end result, 4-5 years further down the line the costs have been recuperated and the ISP is making a killing.

US

[eastmen]

Ahh, but you seem to miss something. With that 20Mbps you should really be getting +4Mbps of true download speeds. That's more that enough. I've downloaded content off the net that are around 8-12GB and it takes 30min. for 8GB at true 4Mbps speeds. So, in my experience I believe that lovely 20Gbps you say you have is already enough for now.

Right but who has the bandwidth to max out a million users trying to download a game on release day ? The only way it coudl work is by using a bit torrent setup. However some isps (comcast) have caps and if your downloading and up loading it could cause problems for many users esp around the holidays as if this generation is any example , I will be downloading 6 games in nov. At 15 gigs each that can add up to alot of bandwidth esp if its a bit torrent setup.

It sucks for you guys that you have limits . Fios doesn't limit me and i've once upload 200 gigs of data and downloaded just under 100gigs in a single month of usage ( uploading high def video of a vacation to friends across the country) I would go nuts if i had a hard cap.

[Unknown Soldier]

Thing we have over here is that we get a group together. One person d/l's one thing and another d/l's another and then we get together at lans and share the content. This seems to work quite nicely, but means that you have to wait a bit to get the content.

Waiting a few days for the release of a new game is not a problem, patience is a virtue.

US

[specwarGP2]

Eastmen,

To achieve your proposed 240MBps transfer rate you need to stripe 8 flash devices. Even if a 32GB SDHC card costs only $1.00 (lol) by 2010, that's $8.00 + control logic, and the total capacity is still only 32GB. That is an totally unecessary cost a publisher should burden.

Do you realize that a 7200rpm HD has a sustained transfer rate of about 55MBps - 60MBps? Does that seem slow to you? If BD drives are not fast enough to stream textures, then copy files to the hard drive, just like its done now. Except by 2010 terabyte size HDs will be commonplace, and installing games will be less of an issue due to space constraints.

Finally, the examples you've used to illustrate why we need a 240MBps transfer rate really have nothing to do with transfer rate at all. Crappy textures in GTA4 is a result of not enough system RAM. Your COD example is flawed as well. the effects of an explosion in an environment need to be visible immediately, therefore those textures need to reside in RAM as well.

[Gubbi]

Some bad math in this thread...

and it takes 30min. for 8GB at true 4Mbps speeds.

4Mbps = 500KB/s = 30MB/minute = 1.8GB/hour

Originally Posted by specwarGP2

Eastmen,

To achieve your proposed 240MBps transfer rate you need to stripe 8 flash devices. Even if a 32GB SDHC card costs only $1.00 (lol) by 2010, that's $8.00 + control logic, and the total capacity is still only 32GB. That is an totally unecessary cost a publisher should burden.

8 x 32 GB = 256GB

Cheers

[aaronspink]

Originally Posted by eastmen

Doesn't bluray require extra steps including that coat of anti scratch resistant stuff they put on it which would keep it more expensive than dvd ?

it adds minor cost but not as much as requiring 2x pressings. Right now the cheapest per GB is a single layer HD-DVD cause it has basically the exact same pressing costs as a single layer DVD. BR is currently still 2-3x per layer over DVD but a lot of that has to do with capacity and depreciation both of which will even out over the next 3-4 years. BR will always be more expensive than dvd simply because of the machine time but it should come down to within 10% over time.

[Shifty Geezer]

Originally Posted by Gubbi

8 x 32 GB = 256GB

Cheers

SpecwarGP2 is seeing parallel-access flash storage as similar ot a RAID striped HDD array - you only gain higher transfer at a loss of overall HDD capacity. Carts with parallel data supply will presumably need to be similar striped with duplicate data, or instead would need to supply different parallel bits of data at the same time, which is an IO headache and would be prone to severe bottlenecks when one chip has all four textures being requested.

[Gubbi]

Originally Posted by Shifty Geezer

SpecwarGP2 is seeing parallel-access flash storage as similar ot a RAID striped HDD array - you only gain higher transfer at a loss of overall HDD capacity.

No you don't. Think Raid-0

Cheers

[Shifty Geezer]

You're right. Never using Raid I forget all the options, but redundancy is ther for security, not performance. Raid-0 increases performance at the same HDD capacity.

[Unknown Soldier]

Originally Posted by Gubbi

Some bad math in this thread...

4MBps = 500KB/s = 30MB/minute = 1.8GB/hour

Cheers

Erm, i'm talking true 4Mbps throughput not a 4Mbps Line.

and it takes 30min. for 8GB at true 4Mbps speeds.

That's 8000/4/60=33.3min.

And yes, i've done it so I know it's real.

US

[Gubbi]

Originally Posted by Shifty Geezer

Raid-0 increases performance at the same HDD capacity.

Same capacity ? If you have N devices of M capacity and O bandwidth you get N x M capacity *and N x O bandwidth with RAID-0.

Cheers

[Gubbi]

Originally Posted by Unknown Soldier

Erm, i'm talking true 4Mbps throughput not a 4Mbps Line.

That's 8000/4/60=26.6min.

And yes, i've done it so I know it's real.

US

Erh, you're measuring true megabits per second as opposed to what? false ?

I'm guessing you're mixing up your bits and your bytes. 4MBps throughput would require 32Mbps of bandwidth (plus bandwidth for communication overhead).

Cheers

[Unknown Soldier]

Originally Posted by Gubbi

Erh, you're measuring true megabits per second as opposed to what? false ?

I'm guessing you're mixing up your bits and your bytes. 4MBps throughput would require 32Mbps of bandwidth (plus bandwidth for communication overhead).

Cheers

Erm, yes sorry, 4MB/s

But as I've said, it's certainly achievable in 33min.

And as for people with 20Mbps, which is a true 2.50 MB/s throughput, an 8GB download will take a meesly 53min.

Not bad imo.

US

[Gubbi]

Originally Posted by Unknown Soldier

And as for people with 20Mbps, which is a true 2.50 MB/s throughput, an 8GB download will take a meesly 53min.

Not bad imo.

Certainly isn't. I also think we'll see a Steam-like download service next gen (downloads and disc-based). If people pre-order (and pay) a significant amount of game data can be pre-distributed to the end-user (similar to what Valve has done) and thus lower contention on release-day.

Cheers

[Shifty Geezer]

Originally Posted by Gubbi

Same capacity ?

Yes, total HDD capacity.

[Squilliam]

Perhaps if we looked at it from another perspective in terms of cost/benifits?

If a console in 2012 cost $350 to manufacture with an optical drive and $300 without, couldn't the console manufacturer substitute the extra profit per unit for reduced royalties per game copy? Currently there are royalties of ~$10 being paid by publishers per game disk sold, if that was reduced to $5 it would be quite sufficient to offset the increased cost of a flash based alternative wouldnt it?

In addition to this, it would be better for a console manufacturer after the current generation to design and build a console with as low cost as possible and with the highest marketability possible. By removing the optical disk component it would allow the console to be packaged into a wider variety of 3rd party devices such as TVs, Cars, Cable devices, all in one entertainment systems (Stereo/Movie/Gaming) and enable these devices to leverage the HD consoles current and future improved multimedia abilities.

[Acert93]

Originally Posted by Gubbi

Certainly isn't. I also think we'll see a Steam-like download service next gen (downloads and disc-based). If people pre-order (and pay) a significant amount of game data can be pre-distributed to the end-user (similar to what Valve has done) and thus lower contention on release-day.

Cheers

That has been my guess.

Likewise a flash memory cache and HDD seem like the easiest / cheapest combo to keep costs low, offset some of the optical transfer speed issues, and tier console pricing (via SKUs). The HDD is finding itself more and more useful this generation (unlike the original Xbox) and with the increasing online activity of users, DLC, and opportunities for post-release revenue streams that the HDD opens up I think mass storage of some kind will be an essential part of new consoles from every maker (it already is this general for all practical puproses, even the Wii and 360 Arcade have some basic standard mass storage). The real question is "what type" and I think a flash cache / HDD combo has a solid performance envolope with a known economic cost. Unless someone can wow Sony, MS, and Ninny with some cheap / high performance SSD technology there is little point. A flash/HDD combo in the hundreds of GBs can be projected to be under $50 cost in 2011. Although Digital Distribition is beginning to weigh into the discussion, it isn't mainstream or world wide and I think people who popo the significance of a 250GB cap on monthly bandwidth are in for a rude awakening when they see how much active gamers/media consumers use per month (note: I have been on the IT end of documenting user bandwidth consumption and no one ever thinks they use what they do). Hence optical media for distribution is a given: it is cheap, it is big, and there are established outlets. Issues are a known quanitity that can be addressed in hardware and software design.

No other media distribution has the same strength of consumer access and low production costs. The problem is speed; game distribution on expensive ROMs/Flash isn't the solution, and DD doesn't have the consumer reach. A large HDD cache, ala traditional PC gaming, solves a bit of this problem and the HDD resource can be leveraged for other technical and fiscal purposes.

[Acert93]

Originally Posted by Squilliam

If a console in 2012 cost $350 to manufacture with an optical drive and $300 without, couldn't the console manufacturer substitute the extra profit per unit for reduced royalties per game copy? Currently there are royalties of ~$10 being paid by publishers per game disk sold, if that was reduced to $5 it would be quite sufficient to offset the increased cost of a flash based alternative wouldnt it?

I am not sure your numbers add up though. DD lacks reach (loss of $, customer base potential) and solid state technology for distribution of games is going to be more than the $5 margin -- plus, VERY importantly, if you produce 1M copies of a game and sell 500k you just ate a ton of your profits. This is one of the reasons people HATED the N64. Even if you could get your flash disks down to $5 manufacturing cost, that is still almost $5 MORE than an optical disk. And at 500k units you are looking at millions and millions of dollars of unused inventory sitting around. And of course what do the console makers have to gain? The average consumer buys ~11-13 games a generation per console (historically), so you have traded negatively, increased risk through inventory, etc. And I dare say $5 per flash disk is under estimated, and $50 for an optical drive next gen is over estimated. A 30-50GB game is going to cost a pretty penny in flash costs and a basic BDR is going to cost much less than $50, especially when sales pick up in 2013-2016.

Your proposal has a lot of risks for everyone involved to break even, it is a lateral move. This is where I think optical / HDD combo works better as it has lower cost and more vertical profit potential.

In addition to this, it would be better for a console manufacturer after the current generation to design and build a console with as low cost as possible and with the highest marketability possible. By removing the optical disk component it would allow the console to be packaged into a wider variety of 3rd party devices such as TVs, Cars, Cable devices, all in one entertainment systems (Stereo/Movie/Gaming) and enable these devices to leverage the HD consoles current and future improved multimedia abilities.

The question is market size: a lot of these are niche, unproven, or smaller markets for consoles compared to the relevance of the "plays Blu-ray" bulletpoint. HD optical is still in the early stages of amassing marketshare, but by 2011 the HD display and HD media market should be quite substantial and people will factor this in. Even if not a sales point, lacking such can be a demerit.

[Squilliam]

@ Joshua Juna - yeah you've got me there!

[MfA]

Originally Posted by eastmen

You have any data for the costs of replecation ?

Google for : dvd replication prices