What Will Drive Future Demand for Bandwidth?

[west]

Recently ItsAValueTrap posted a great article on his blog about future demand for different technologies (e.g., CPU power, network bandwidth, etc.)  Please read it here:

 

http://glennchan.wordpress.com/2014/10/04/what-happens-when-technology-is-good-enough/

 

I would summarize, but by summarizing it I think I would skew what he says in a way that more reflects what my perspective is.  So I'll leave it for you to read instead of summarizing.  :D

 

One of the things he talks about is how he thinks that the future demand for bandwidth will decrease once there is full 4K penetration in the cable and other video delivery businesses in the US.  I disagreed with him on this, or at least I felt that assuming demand for more speed/bandwidth would taper off is a dangerous assumption.

 

My feeling on the matter is that there always seems to be something new in technology that creates demand for more.  While this isn't true in the realm of CPU power (most people have enough CPU power at this point to meet their needs) or in say sound cards (when's the last time anyone here worried about what sound card is in their computer?), it is true for almost every other domain in technology.  For example, the more hard drive space is available to consumers, the more people seem to need more hard drive space.

 

My feeling is that while there is a chance that demand could taper off at some level, we are a long ways away from that point.  There could be more devices on the network, whether at home or at work.  There could be more applications, some of which we haven't even heard or thought of.  (Holographic video calls anyone?)  Who knows.  I just think we're a long, long way off from satisfying any current or future complete bandwidth demand.

 

I think Glenn's thinking goes something along the line of: Most internet traffic is driven by video and most people don't care too much about 4K video (let alone 8K video).  Even if they do care about 4K video, most networks will be able to handle it within a few years.  Then what's left to drive bandwidth?

 

(Correct me if I'm wrong on this Glenn.)

 

What do you guys (preferably you guys who have knowledge on the matter) think?

 

[ItsAValueTrap]

I think Glenn's thinking goes something along the line of: Most internet traffic is driven by video and most people don't care too much about 4K video (let alone 8K video).  Even if they do care about 4K video, most networks will be able to handle it within a few years.  Then what's left to drive bandwidth?

Yep, pretty much!

 

As far as storage goes, I think we're also close to the point of being good enough for most people.  People can (A) switch from hard drives to SSDs and (B) backup their data via the cloud because the demand for storage hasn't grown that much relative to advances in technology.

 

In the past, something new would always show up and drive demand for better CPUs, more storage, more RAM, more bandwidth, etc.  But I think this is happening less and less nowadays.  I'm sure a lot of offices are still running Windows XP.  Here are some 10-Ks that talk about Windows XP or2000:

http://www.sec.gov/Archives/edgar/data/708818/000070881814000008/qsii-10k_33114.htm

http://www.sec.gov/Archives/edgar/data/945384/000135448814001527/cicn_10k.htm

http://www.sec.gov/Archives/edgar/data/1185348/000118534814000002/praa-20131231x10k.htm - PRAA - "Our desktop PCs run the Windows XP or Windows 7 operating system."  (No Vista.)

 

2- Adding 3-D to video might require 10-30% additional bandwidth.  (Don't take that number too seriously.)  Increasing the frame rate also doesn't require a lot of additional bandwidth due to the magic of video compression.

 

I'm not sure what you can do beyond that.  There's no killer consumer app that is driving demand for bandwidth other than video-related applications.

[SpecOps]

Speaking personally, I want bandwidth that allows me to download film/tv in less than 10 minutes. Streaming is good, but there are some things I prefer to download and save.

 

At the moment there is a lot of lossy compression that goes into films, even HD, in order to get them small enough for download and storage. When I rip a Bluray it can be well over 25Gb, storing my large collection electronically isn't possible without compressing them and losing quality. I would ideally have my whole collection electronic and do away with discs altogether, I hate having 'things' cluttering the house.

 

On processors, I was shocked recently when my office upgraded our computers to ones with Intel Core i5. I bought a laptop with that processor myself over 5 years ago, and after searching realised that the Intel Core iX range are still being sold as standard.

[prevalou]

for those interested

 

http://www.libertyglobal.com/pdf/public-policy/Liberty-Global-2014-Future-Of-The-Internet.pdf

 

Bandwidth is not the only parameter

[StubbleJumper]

"640k ought to be enough for anybody"  - Bill Gates

[oddballstocks]

I think Glenn's thinking goes something along the line of: Most internet traffic is driven by video and most people don't care too much about 4K video (let alone 8K video).  Even if they do care about 4K video, most networks will be able to handle it within a few years.  Then what's left to drive bandwidth?

Yep, pretty much!

 

As far as storage goes, I think we're also close to the point of being good enough for most people.  People can (A) switch from hard drives to SSDs and (B) backup their data via the cloud because the demand for storage hasn't grown that much relative to advances in technology.

 

In the past, something new would always show up and drive demand for better CPUs, more storage, more RAM, more bandwidth, etc.  But I think this is happening less and less nowadays.  I'm sure a lot of offices are still running Windows XP.  Here are some 10-Ks that talk about Windows XP or2000:

http://www.sec.gov/Archives/edgar/data/708818/000070881814000008/qsii-10k_33114.htm

http://www.sec.gov/Archives/edgar/data/945384/000135448814001527/cicn_10k.htm

http://www.sec.gov/Archives/edgar/data/1185348/000118534814000002/praa-20131231x10k.htm - PRAA - "Our desktop PCs run the Windows XP or Windows 7 operating system."  (No Vista.)

 

2- Adding 3-D to video might require 10-30% additional bandwidth.  (Don't take that number too seriously.)  Increasing the frame rate also doesn't require a lot of additional bandwidth due to the magic of video compression.

 

I'm not sure what you can do beyond that.  There's no killer consumer app that is driving demand for bandwidth other than video-related applications.

 

Conventional storage yes, you can buy a 4TB drive at Costco for sub $100, and that includes an external enclosure.  But SSD is a different story, I'm really looking forward to price drops with that.  Computers are going to really speed up when SSD is standard.  The largest limiter to performance is hitting the disk right now.  Taking an old computer from a conventional drive to SSD breathes new life into the machine.  This needs to happen across the board.

[yadayada]

Gotta look at price too. Sure ill prefer 4k above 2k, but you need a lot more bandwith. And I only care about it a little bit more as you can barely see the difference. I wouldn't like paying twice as much for it. I fail to see how for example people cutting the cable will suddenly pay more for internet. Unless you throttle the speed like they do now. But if you already have some of the slowest internet in the western world, at some point people will not take that anymore.

 

And at some point services like google will lay down their own fibre if things like youtube get large enough?

 

[Ross812]

I think this is off the mark for both storage and bandwidth. Companies providing storage and backup (i.e. the cloud) still have to buy HDDs. These companies also have redundant arrays, something that a lot of individuals do not have the money/aptitude to setup. So and individual storing 700GBs on a 1TB drive at home is now storing 700 to 2.1TB in the cloud (mirrored 1 or two times). Projected storage needed in the future:

 

http://www.cisco.com/c/dam/en/us/solutions/collateral/service-provider/global-cloud-index-gci/Cloud_Index_White_Paper.doc/_jcr_content/renditions/Cloud_Index_White_Paper-02.jpg

 

Point and shoot cameras now have wifi for easy backup, smartphones, tablets… The cloud will keep growing exponentially as cloud backup and sharing gets easier and easier. Is You-Tube growing?

 

I don’t know what is going to happen with bandwidth consumption after the majority of people are watching 4k videos on Netflix. I would not bet against further innovation in the entertainment segment. “4K is all we will every need” sounds like it’s good for 5-10 years. Also, don’t underestimate the multitude of devices that are coming online. You can now buy a washer/dryer, thermostat, lighting system, scale, printer, and refrigerator that are wifi connected. These devices may not use the bandwidth of 1 HD movie in a year together, but if everyone has these devices in the future... Smartphones are also using more and more data year by year as websites become more complex and sharing has become easier. 

 

[anders]

If the discussion is about data speed and storage i think we are just seeing the beginning ;D

 

I think below are nr of things that will continue to grow many decades to come thanks to generation Z, that will require much higher bandwidth:

 

-Cloud computing: Probably more than half of all IT will be in the cloud within 10 years and asia-pacific will most likely in a couple of years outpace north america in global cloud computing usage.

-Artificiell intelligence: Googles recent purchase

-Mobility: Photos, music, email

-Nanotech, 3D printing, Smart-phones/cars/boats/planes, Smart-homes/buildings/citites, wearable tech

-VoIP: Microsoft recent purchase of Skype.. integrate that with outlook and we have conferences, voice emails.

-Gaming: Virtual reality hardware (oculus and playstation morpheus ready to hit the shelves soon), potential to recreate matrix

-Space Tech: googles purchase of titan aerospace

 

And yes, I believe people are going to be satisfied with what they have until they see the new thing the neighbour just bought...  ;)

 

Rgds

 

 

[rkbabang]

Like anything else, there is a difference between what you "need" and what you'd like.  I'd love it if I could back up all of my computers and devices to the cloud in about 2 seconds (maybe 6TB of data) and restore any one of them in a second.  But I certainly don't "need" that kind of bandwidth.  As far as storage and CPU/GPU speed.  I'd love to be able to have that 6TB of data on redundant SSDs.  rather than HDDs.  I'd love to be able to render 2 hours of 4K video in under a minute.  None of these things are possible right now.  I'm converting my old home movies from tapes to only standard-def video and for a 1.5 hour tape it takes over 2 hours to do that on my 8-core 3.2GHz machine with 16GB of RAM.  I have to play the tape (about 1.5 Hours) to get it into the software (Adobe Premiere) then it takes about 30-45 min to render it and save it to the format of my choice.  I have probably 60 of these tapes to do and I'm on tape 12 right now.  It is a long process. If this was 4K video it would be frustrating as hell.

 

Asking how much bandwidth, or CPU speed, or storage do you need is like asking how much horsepower do cars need.  Certainly we had enough 50 years ago, but you can buy 400-800HP cars now, hell there are 300+ HP 6 cylinders now.

[Guest longinvestor]

What will drive future BW? Things we don't know of today.

 

Data transport is the third(last) leg of the data stool. The first two are processing speed and data storage. These have already played out. BW is simply following.

[west]

What will drive future BW? Things we don't know of today.

 

I think this sums up my thinking nicely.  There is rational argument that can be made that we're nearing satisfying bandwidth demand based on current needs.  However, when it comes to technology, it's very dangerous to assume that tomorrow's current needs will be the same as today's.  ("This time is different" anyone?)  While it may end up being true, I wouldn't bet on it (and I'm guessing if I had to put a bet on one side of the argument or the other, my bet would probably be on the other side).

 

Two potential reasons why bandwidth demands could increase:

 

1. IPv6 allows the following number of devices to exist on the internet:

 

340,282,366,920,938,463,463,374,607,431,768,211,456

 

That is, 667 sextillion devices per square meter on earth.  The world is upgrading to IPv6 because the 3.7 billion addresses provided IPv4 wasn't enough.  And those are just the addresses for the public network.

 

If device density increases, the bandwidth demanded per consumer location will increase, which will increase the demand for more bandwidth on broadband networks.

 

2. Holographic (not just 3d) video chatting.  Maybe not now, or in ten years, but I have a hunch it will exist in most of our lifetimes.  This may sound like it's straight from out of Star Trek, but honestly it's not that complex on top of what we already can do (video encoding, 3D scanning, etc).  It wasn't too long ago that having a device in your pocket that let you talk to anyone in the world at the touch of a button (let alone have access to pretty much the entirety of the world's knowledge) would have been considered sci-fi as well...

 

Maybe neither of the two things mentioned above will drive future demand for bandwidth that much.  However, I bet there will be something.

[Guest longinvestor]

What will drive future BW? Things we don't know of today.

 

Data transport is the third(last) leg of the data stool. The first two are processing speed and data storage. These have already played out. BW is simply following.

 

IMO, it is not the cool & sexy aspect of tech that gets adopted, rather the utility value of a tech. For ex. Cheap storage (memory)brought on the digital camera revolution. Cheap memory helped photography overcome the cost drag of film. How many of us chose to ration the number of pictures taken because we did not have an extra film or to avoid running up film costs?

 

BW has got to be similar. Cost needs to keep falling and we will see new things spawn. My favorite application already enabled by cheaper BW is Voice-over-IP. Twenty five years ago, we routinely paid >$500 per month because of long distance and international calling on our home phone. Now, we have all-you-can eat home calling for <$25 per month. Netflix is standing proof of cheaper video transport, enabling cord cutting. Mobile devices are going the same way, albeit not fast enough in the US. Check out what's going on in Japan, lead by Softbank. Video conferencing will likely cut into unnecessary business related air travel.

 

[ItsAValueTrap]

I'm asking about what will happen to shareholders.

 

If you look back at history, investors got waaaaay too excited about Nortel Networks and their voice over IP(-like) technology.  That was partly due to fraud.  But it also has to do with VOIP becoming good enough.  Only so much bandwidth is needed for VOIP.  We reached that stage a long, long time ago.  Actually since we reached that stage the quality of calls has gone down due to people using calling cards and being ok with cost-cutting measures (which result in audio compression artifacts, latency/lag, etc.).

 

Eventually, maybe the growth in Internet traffic will stop.  I certainly think that Netflix and other on-demand services will drive huge demand for Internet bandwidth in the next few years.  But that might stop.  And if that happens, it may not be a good thing for Cisco, Level Three, Akamai, etc.

 

Intel is seeing the revenues from its PC Client group drop (even though more people worldwide are using computers).

 

"640k ought to be enough for anybody"  - Bill Gates

Apparently he never said that.

[merkhet]

I suspect that if we had good enough and consistent enough bandwidth, our computers, tablets, etc. could become terminals such that we begin to stream things that were once localized.

 

That would probably require substantial bandwidth capacity.

[ItsAValueTrap]

I suspect that if we had good enough and consistent enough bandwidth, our computers, tablets, etc. could become terminals such that we begin to stream things that were once localized.

 

That would probably require substantial bandwidth capacity.

 

We've been doing that for audio and video.  On the audio side you have Pandora, Grooveshark, Spotify, etc.  On the video side you have Netflix.

 

But the thing about streaming audio and video is that there are limits to how much bandwidth we need.  For audio we don't need quality that is better than 192kbps mp3s.  For video we don't need better than 4K.  Once we get close to the limits of human perception, the demand growth for bandwidth stops.

[writser]

Having studied computer science myself I enjoy the scientific side of this discussion. However, we nerds have a tendency to get caught up a bit in the technical details (because we love them). So we talk about the bandwidth requirements of 4k video, IPV6, and 3d video chatting.

 

But as an investor, I think trying to estimate the future demand for bandwidth is not that useful. First of all: in tech it is notoriously hard (impossible) to predict the future. Maybe ADSL or wireless will completely wipe out cable. Maybe new technologies will require much more / much less bandwidth in a decade. Who knows? Secondly, is it a given that increased demand for bandwidth leads to higher profitability? I don't think so. Maybe it's better if the industry stagnates and consolidates. Also, another important driver for the future: growth in emerging markets, has not been mentioned in this discussion so far. Not to mention the regulatory landscape. And not to mention the prospects currently priced in in individual stocks.

 

I think a lot of other questions are more relevant from an investors perspective. Not only is it impossible to determine the future demand for bandwidth, it is also impossible to determine what effects this will have on the industry.

 

But by all means please continue the discussion :) .

[peter1234]

not everything that can be counted counts, and

 

not everything that counts can be counted.

 

(Attributed to Einstein)

 

;)

[adesigar]

I suspect that if we had good enough and consistent enough bandwidth, our computers, tablets, etc. could become terminals such that we begin to stream things that were once localized.

 

That would probably require substantial bandwidth capacity.

 

We've been doing that for audio and video.  On the audio side you have Pandora, Grooveshark, Spotify, etc.  On the video side you have Netflix.

 

But the thing about streaming audio and video is that there are limits to how much bandwidth we need.  For audio we don't need quality that is better than 192kbps mp3s.  For video we don't need better than 4K.  Once we get close to the limits of human perception, the demand growth for bandwidth stops.

 

I don't think he is talking about streaming Video and Audio. What I think might happen (and its probably what merkhet is talking about) is all our devices are like Chromebooks but more than just Data is in the cloud. We have an interface and the CPU/GPU/RAM/HDD/VideoCard/OS/Software(Office/Databases/Games) is all in the cloud. No need to maintain hardware or upgrade it. Interface separates data storage from Hardware layer. Rent it by the month switch to a different configuration/OS at anytime.

[merkhet]

I suspect that if we had good enough and consistent enough bandwidth, our computers, tablets, etc. could become terminals such that we begin to stream things that were once localized.

 

That would probably require substantial bandwidth capacity.

 

We've been doing that for audio and video.  On the audio side you have Pandora, Grooveshark, Spotify, etc.  On the video side you have Netflix.

 

But the thing about streaming audio and video is that there are limits to how much bandwidth we need.  For audio we don't need quality that is better than 192kbps mp3s.  For video we don't need better than 4K.  Once we get close to the limits of human perception, the demand growth for bandwidth stops.

 

I don't think he is talking about streaming Video and Audio. What I think might happen (and its probably what merkhet is talking about) is all our devices are like Chromebooks but more than just Data is in the cloud. We have an interface and the CPU/GPU/RAM/HDD/VideoCard/OS/Software(Office/Databases/Games) is all in the cloud. No need to maintain hardware or upgrade it. Interface separates data storage from Hardware layer. Rent it by the month switch to a different configuration/OS at anytime.

 

Yup, that's what I meant -- so it's not so much that we haven't already created dumb terminals (out of our TVs, for sure) -- rather the number of dumb terminals that we have will increase.

[rkbabang]

I suspect that if we had good enough and consistent enough bandwidth, our computers, tablets, etc. could become terminals such that we begin to stream things that were once localized.

 

That would probably require substantial bandwidth capacity.

 

We've been doing that for audio and video.  On the audio side you have Pandora, Grooveshark, Spotify, etc.  On the video side you have Netflix.

 

But the thing about streaming audio and video is that there are limits to how much bandwidth we need.  For audio we don't need quality that is better than 192kbps mp3s.  For video we don't need better than 4K.  Once we get close to the limits of human perception, the demand growth for bandwidth stops.

 

I don't think he is talking about streaming Video and Audio. What I think might happen (and its probably what merkhet is talking about) is all our devices are like Chromebooks but more than just Data is in the cloud. We have an interface and the CPU/GPU/RAM/HDD/VideoCard/OS/Software(Office/Databases/Games) is all in the cloud. No need to maintain hardware or upgrade it. Interface separates data storage from Hardware layer. Rent it by the month switch to a different configuration/OS at anytime.

 

Yup, that's what I meant -- so it's not so much that we haven't already created dumb terminals (out of our TVs, for sure) -- rather the number of dumb terminals that we have will increase.

 

Imagine if devices could move data into/out of  the cloud as quickly as computers today can move data in or out of DDR4 memory, equivalent to moving 64bits at a time at a rate of 3.2GHz.  That is over 200 Gbps, more than 200 times Google Fiber's 1Gbps.  Now imagine 15-30 devices in every house (hundreds to thousands in every office building) all trying to do that at the same time.

 

[ItsAValueTrap]

Regarding dumb terminals over the Internet...

 

I don't think they will take off.  But suppose that they do.  Maybe one day cloud gaming will work.  I'm pretty sure that video will max out at 4K.  We already know that moviegoers don't care about 4K and are satisfied with 2K.

 

Once the demand for bandwidth growth stops, some tech companies may face headwinds.  Historically, what has happened is that revenues shrink.

[JAllen]

I strongly disagree that bandwidth requirements are stagnating or will stagnate. 

 

 

There ARE consumer apps that are consuming exponentially more bandwidth-storage (you can't really have one without the other).  GoPro just released a 4K camera.  This is a prime example of an app that will result in much greater bandwidth consumption.  While 4K TVs and monitors are mostly still expensive, they will decline in price just like every other increased screen resolution has.

 

 

http://arstechnica.com/business/2012/05/bandwidth-explosion-as-internet-use-soars-can-bottlenecks-be-averted/

 

 

Voice calls just started going to LTE = more bandwidth.

 

 

Still 4 billion people are not using the Internet at all.  But they will come online over the next two days.  This will equal massively more bandwidth.

 

 

And if we weren't already pushing up against bandwidth supply, indicating growing bandwidth, I could potentially agree.

 

 

We will always demand faster/more bandwidth.  It still takes me an hour to download movies with Comcast.  I want more and faster bandwidth.  I would pay serious money to have Gigabit speeds, but alas, they aren't even available.  I can't wait for that.

 

 

Until we don't have to ever wait for downloads we will want more and faster bandwidth.

 

 

Check out the bandwidth chart here: http://perspectives.mvdirona.com/default,date,2008-12-02.aspx

 

 

I often have to wait for Netflix or Amazon Instant Video to load my video.  Bandwidth isn't even close to slowing down.

 

 

There are still tens of millions of Americans on DSL getting 1-2 Mbps.  When/if they get true broadband or better yet Gigabit speeds, will create another tailwind for more bandwidth consumption.

 

[CorpRaider]

Three words:  Virtual Reality.....Porn.

[adesigar]

Regarding dumb terminals over the Internet...

 

I don't think they will take off.  But suppose that they do.  Maybe one day cloud gaming will work.  I'm pretty sure that video will max out at 4K.  We already know that moviegoers don't care about 4K and are satisfied with 2K.

 

Once the demand for bandwidth growth stops, some tech companies may face headwinds.  Historically, what has happened is that revenues shrink.

 

Yes but you are comparing 4K on a single screen for a 2hr movie. People using cloud Pcs would use multiples of that. I use 4 screens for 10+ hrs a day. I know companies who have bought racks at hosting companies and moved their servers offsite because it was cheaper to pay for hosting/bandwidth than to pay for office space, electricity etc.