The closest analogy therefore would be if you left your door open at home and someone - unbeknowst to the house owner - stood outside and told people they were "free to go in and take what they like". This act would not excuse anyone who then took it upon themselves to steal from your house.
That's a pretty bad analogy since you can usually tell the difference between a private residence and a public place from looking at it. This is not the case with open access points. If *all* houses looked like pubs and one had someone standing outside (unbeknowst to the owner) saying "come in for some beer", would you have any reason not to assume it was a real pub inviting people in for some beer?
Furthermore I don't think it is "fair" to simply say "well it's the responsibility of the network owner to secure their equipment, if they don't it's fair game".
The problem isn't that network owners aren't securing their access points - the problem is that it's not possible to tell the difference between an intentionally free access point and a badly configured private access point. So it seems we have two choices: 1. Make it illegal to connect to any open access point 2. Make is completely legal to connect to any open access point
There is no middle ground since it's often not possible to tell whether an AP is supposed to be open or not. I'd say that (1) would be a pretty bad state of affairs since there are a lot of legitimately open APs. It'd put the hotspot providers out of business for one, and kill all the free hotspots provided by pubs, cafes, etc.
I have a water tap in my front yard... should you assume that, without asking, you have a right to plug a hose up to it and use as much water as possible?
If you have a sprinkler in your front garden, and it's over-spraying onto my garden, should I be arrested for "stealing" your water?
It's quite clear that almost no one would want someone stealing their bandwidth for free if they knew what bandwidth was all about and the consequences, so respect that and don't use it.
That's funny - quite a few of the pubs around here intentionally set up open access points for their patrons to use. How are you suggesting I tell the difference between an intentionally open AP and an accidentally open AP?
Do not use unless you know for sure that you are allowed. An SSID of "UseThisWiFi" or similar.
It's not always clear what is public or not without explicit knowledge of what SSIDs the public access providers use.
Lets take an example. You come across an access point called "MyCloud" - there is nothing in that name that tells you it's a public access point if you didn't already know that the hotspot provider called "The Cloud" happened to exist and use "MyCloud" as their SSID.
Another example. You're at an airport and you find an open access point called "AirportWIFI", or you're at a pub and find "PubWIFI" - are you to assume that these are intentionally open for the use of the airport or pub patrons, or are they private networks that have been accidentally left open.
Another example. BT runs 802.11 hotspots, but they also provide 802.11 equipment to people subscribing to BT's ADSL connections. You find an open access point called "BT Voyager" - are you to assume that it's one of their hotspots or a private access point that's been misconfigured?
The fact is that there is no way to tell the difference between a public free hotspot and a misconfigured private access point. It seems to me that if an access point advertises itself as an open AP, and when you conntect it leases you an IP address and then gives you unrestricted access to the internet then you should be well within your rights to assume it's a public hotspot since there is no way to tell any different. On the other hand, if someone has made some attempt to secure it, even if it's just hiding the SSID, you should assume it's not intended for public use and leave it alone.
Not only that, but if your home access point is called "Linksys" then your laptop will probably quite happilly associate with any other open access point with the SSID of "Linksys" even if you don't ask it to.
The US has fairly clear and well enforced truth in advertising laws. Doesn't the UK?
Yep, and I don't really know why the Advertising Standards Authority doesn't come down on some of these ISPs. Although Orange has recently bee slapped down for making claims about "unlimited broadband" without mentioning a fair use policy. The ASA seems fairly happy with people saying "unlimited* broadband" and then having small print stating that fair use policy applies.
This seems wrong to me - if there is a fair use policy that places limits on how much you can download then it clearly isn't unlimited. Especially since those fair use limits actually do affect a significant number of people. It also makes it hard to compare what ISP is going to be best for you if they all hide their limits away in the small print (or worse: don't publish their limits at all).
Also, some of the fair use policies seem to place limits on much more than fair use - for example, Orange's "unlimited" GPRS tariff has a fair use policy which explicitly bans the use for instant messaging and VoIP. I wouldn't call that fair use since it won't impact other users of the system - it's just Orange preventing you from using services that would undermine their telephony business.
It seems companies can make untrue claims about their products in the UK and not get sued?
The UK culture is nowhere near as litigeous as the US. For the most part I consider that to be a Good Thing since it helps keep things like insurance costs down because there aren't so many frivolous and money-grabbing lawsuits. However, that gradually seems to be changing as more and more ambulance chasing lawyers seem to keep popping up and avertising "no win no fee" services for all sorts of situations.
False advertising claims are generally handled by the advertising standards authority, and for the most part they do a reasonable job. Although their fines are usually small and frequent enough to prevent companies from _continuing_ to misadvertise rather than big one-off fines which scare companies away from doing it in the first place. Sadly, the ASA sometimes make some bad decisions (as mentioned above), but I guess that's not a huge amount different from launching a class action and having the judge make a bad decision.
I don't _think_ there is such a thing as "class action" in the UK. And the average punter doesn't want to spend the time and money suing - unless they actually lost a lot of money they are more likley to just stop using the company in question and bad-mouth them to their friends.
Not sure if this is the case in the UK, but in the US there always seems to be the same statement--"up to X mbits". I don't know of any large ISPs that do not use that phrase.
Yup, they all use that phrase - I wasn't arguing that that was misleading (although now you come to mention it, if your network is overloaded to the point that there's no way anyone's going to get 8Mbps you probably shouldn't be allowed to advertise "up to 8Mbps").
I do not know any company that explicitly states their residential service is "unlimited".
Thats just from a quick Google for "unlimited broadband". I hear radio ads quite frequently for "unlimited broadband" too, and infact it wasnt that long ago that Tiscali's radio ads said "Up to 8 megabyte unlimited broadband" (which was clearly a mistake and whoever let that through needs a kicking).
Right now in the US, no company has ventured out into metered billing like PlusNet has.
It's pretty common place in the UK amoungst the smaller ISPs. The bigger ones are still misleading everyone with their "unlimited" claims though.
Ultimately, many who complain a lot about residential service not being unlimited have no clue what such a service actually costs.
The customer shouldn't need to care how much the service costs the ISP. If the ISP sold them an "unlimited" then the customer has every right to expect it to be unlimited - if the ISP misadvertised it then it's the ISP's responsibility to swallow any costs required to meet their commitment.
However, the downside is that photons with higher energy than that bandgap, well, the extra energy is essentially wasted.
Since the energy that isn't converted into electricity becomes heat, would it not be sensible to mount the photovoltaic cells on the surface of thermal solar panels (i.e. water heaters)? That way, the solar panels on your roof can provide the power for your TV and the heat for your shower at the same time.
Some Linux updates come with the request for you to restart the affected service (Samba for example)
I'd prefer to restart a single service rather than the whole machine. Generally I'm not going to notice if Apache gets restarted, for example, but I am going to notice if the machine reboots since I'd lose my X session and everything that was running.
kernel updates basically leave you in the situation that modprobe has its difficulties. Being a wimp, I reboot my PC (a desktop) when that happens.
Under Fedora, RHEL and CentOS, yum installs the new kernel whilest leaving the running kernel also installed. This means you can continue running your current kernel with modprobe working and everything until you decide it's time for a reboot. So generally I only bother to reboot for security fixes.
I don't really disagree - as I said, the fact that you are locked in to a specific protocol in order to access a service (which you now seem to consider is true) is probably unfixable, but that doesn't mean it doesn't cause problems.
The original poster said: By 2020, the current internet will have a level of lock-in that makes Windows look disposable. "Faster" and "safer" will have a tough time overcoming "empty".
And I firmly believe he's right (infact, I think that this level of lock-in happened long ago). So much content and services are available over the internet (or more specifically: IPv4) that you are pretty much locked in to using IPv4.
If someone comes out with a new protocol (which has already happened with IPv6) or even a whole new infrastructure, that new technology is going to have real trouble gaining ground for the simple reason that it has no content/services on it - the content/services are still all on IPv4 (or available through both). The mere fact that IPv4 is "good enough" for most uses means that there is little incentive for everyone to upgrade.
Remember that to roll out an "internet replacement" you're going to need to get the server owners, the connectivity providers and the end users to all upgrade. If one of them doesn't feel it's worth their while then noone gets to play with the "internet replacement" even if they want to.
Sure, you can roll out IPv6 on your own network (I did years ago) and run IPv6 and IPv4 in parallel, but the vast majority of your traffic is going to be IPv4 for years to come unless a major reason comes up for the server owners and ISPs to expend the money needed to upgrade.
With IPv6, there are major reasons for upgrading: - Do away with NAT. This is mostly going to affect the end users and server owners since P2P protocols such as VoIP are becoming more popular. Unfortunately, P2P does work through NAT (but is less reliable) so much of the push to upgrade is gone there and then. - IP address shortage. This is possibly the bigger reason. Even with NAT we _are_ going to run out of IPv4 addresses at some point in the future. However, I feel this is going to be very much like the Y2K problem - everyone can see it coming but noone is willing to spend the money to plan for the future. Eventually this problem is going to hit us hard and there will be an "oh shit the sky is falling" moment where everyone will end up spending way more money upgrading than they would've done if they had upgraded earlier.
No one who has lived in my house in 60 years has ever picked up the phone and it not worked.
That is a different experience than those who use this service have.
I'm the last person to defend Skype, but it seems your analysis is flawed. You have compared the experiences of a single household with the experiences of several thousand slashdotters.
Personally, I have had outages on phone lines on several occasions. I can think of at least 2 occasions when my BT line has completely failed and probably around 3 occsions when my NTL line has failed.
I'm not saying VoIP is less reliable than the PSTN (infact I'm sure it is), but I am saying that the basis for your "proof" is completely bogus.
Oh, please! How is that to be considered a lock in? Does my use of SIP prevent me from using XMPP? No it does not.
How does your use of Office prevent you from using Open Office? It doesn't.
How does your use of MSN prevent you from using XMPP? It doesn't.
Yet, both MS Office and MSN are considered cases of vendor lock-in _because_ you are tied to using a specific program/service/protocol in order to access specific content.
Will the XMPP corporation change the protocol so that I'm required to buy their software all over again? No, that will not happen.
XMPP corporation? XMPP is the IETF standard instant messaging protocol - there is no XMPP corporation.
And yes, the XMPP protocol does change and you are required to upgrade your XMPP client if you want to use the new functionality. Just as Office documents change to incorporate new functionality and you are required to upgrade your copy of Office in order to use it.
In the case of IPv4 and IPv6, that doesn't apply. I can migrate any time I want to. The fact that there's not much point until most of the rest of the world does likewise doesn't mean I'm locked in
You can migrate from MS Office at any time you want. The alternatives won't be much use (assuming they can't use Office documents) if you want to exchange documents with the rest of the world, but there's nothing stopping you switching. I really don't see a difference:
* You can change from IPv4 to IPv6 but there's no point since the content is only available over IPv4. * You can change from Office to $other_office_suite but there's no point since the content is only available in Office format. * You can change from MSN to XMPP but there's no point since the people you want to talk to are only on MSN.
All looks pretty similar to me.
I did, and I chose not to address your point because I felt (and still feel) that you were proceeding from a false premise.
Not addressing my point is one thing. However, what you did was ignore my point and willfully assume I believed the exact opposite of what I stated I believed.
[Office] In particular, there are also issues of deliberate breaking of backwards compatibility to force othewise undeeded upgrades
IPv6 breaks backwards compatibility with IPv4 in the same way. Compare: * You want to open an Office document that was saved from a newer version of Office, thus you have to upgrade your version of Office too. But all your old documents still work just fine with your existing Office. * You want to use a service that is only available over IPv6, thus you have to upgrade your IP stack. But all the old services still work fine over IPv4*
In both cases, you are not _required_ to upgrade, however you do need to upgrade *IFF* you want to access new stuff.
(* as mentioned, this isn't necessarilly the case - if IPv6 takes off then IPv4 services will be erroded and replaced with IPv6-only services. In a way, it makes it _worse_ than Office since you can nolonger guarantee your existing stuff will continue working without an upgrade).
This I do not accept as lock in. There is no key holder here and no potential for abuse
As I have stated before, I do not believe the potential for abuse is a defining feature of lock-in. Sure, if you are locked in then there may be an increased potential for abuse, but that doesn't mean that all lock-in scenarios can be abused.
So I still think you're attempting to trivialise a serious industry problem by conflating it with a completely unrelated problem on the one hand
I'm not intending to trivialise it at all. There _is_ a serious industry problem with lock-in, and that is a big problem. But to believe that there is no such thing as lock-in (and no disadvantages) in situations where a single organisation isn't in a position to abuse people seems extremely naive.
Rather than trying to trivialise the "vendor lock-in" problem, I am trying t
correct... if no firewall is present... if a firewall block UDP packet, then relying is a must...
Or the user could get a clue and poke the correct holes in the firewall. I'm not convinced that engineering protocols to try and bypass the user's own firewall is a bright idea - firewalls are there for a reason.
EVERY other software distribution has [frequent, but not necessarily monthly] updates that require a restart like this.
Umm.. no... My linux machines very rarely need to be rebooted in order to receive security updates - the security updates quietly happen nightly and I rarely need to bother about them. The Windows machines, on the other hand, seem to need a reboot every month.
Basically the problem can be put down to the combination of 2 problems: 1. A operating system with a flawed update system that requires regular reboots for all users at roughly the same time (Windows Update). 2. A flawed protocol that requires all users access the same cluster of centralised servers (Skype).
The solution: avoid using such flawed systems - there are better alternatives available.
And this would be a perfect example of why centralised networks are a Stupid Idea.
Use a decentralised technology such as SIP - if something catestrophic happens, only a few users are affected rather than taking out the whole network and if your service provider provides a sucky quality of service you can go move to a different one (infact you don't even need a server if you're not having to negotiate things like NAT - SIP runs quite happilly as a peer-to-peer application).
You provide a false choice by suggesting ISPs could possibly choose to sell service without any oversubsription. Say the 10mbit service my company offers were sold assuming 100% usage 100% all the time. Depending on infrastructure costs and typical bandwidth costs for a decent size ISP in the US, you would be talking around $500 to $1000 per month. Say you have 1,000,000 customers in a modest size city. Are you suggesting that the ISP should plan on delivering 10 Petabits of data to that community? I'm not sure that that kind of bandwidth is even feasible with current technologies.
If they want to sell "unlimited" connections then yes, they need to provide enough bandwidth to support what they are selling and set their prices appropriately to cover the cost.
Alternatively a more sensible suggestion would be to stop misadvertising their services as "unlimited" and instead include information as to what the limits actually are. Also, providing several different products with different limits and appropriate pricing seems a good idea to keep both the light users and the heavier users happy, rather than having one group of users subsidising the other.
Really, if you advertise a service and use the fact that it's "unlimited" as a selling point, you damned well can't complain when people try to use it in an unlimited way.
An ISP could just change their terms of service and next month the customer can decide if they want to stay.
Herein lies the problem - when everyone else in the market is advertising "unlimited" services, your service doesn't _look_ attactive if it's similarly priced but has published limits even though the "unlimited" services probably have similar limits which just aren't published.
Changing your terms of service also leads to bad publicity (those few unprofitable heavy users can make a lot of noise - I saw it happen when PlusNet changed their terms in this way). And even the less well informed people who don't overuse the service may feel hard done by since you appear to be taking something away that they paid for (even though they aren't using it and their service will probably actually _improve_ as a result).
So in a market where everyone else is misrepresenting their products it's pretty hard to be truthful whilest still looking competetive.
Thankfully here in the UK, many of the smaller ISPs have stopped advertising their connections as "unlimited" and in a number of cases have put together a good explanation of why no ISP can offer unlimited internet connections. I would be interested in how many potential customers ignore that and still believe an "unlimited" connection is better value for money so go elsewhere though.
Not only is the problem not inconveniencing anyone sane
Completely untrue.
If you're using SIP as your instant messaging protocol and I'm using XMPP, one of us has to change protocol (or use both) - that is inconvenient.
Similarly, I have some servers that are only accessible over IPv6 (because I don't have enough public IPv4 addresses to go around). These are specifically servers that I only need to access from locations that I know have IPv6 connectivity so it isn't a problem, but it does mean that if I needed someone with only IPv4 access to connect to them I'd have to jump through some hoops - that is inconvenient.
By your argument a lock-in is a lock-in, and therefore that's a Bad Thing, or at least as bad as any other lock in.
That is explicitly _not_ what I said. If you had read my post you would've seen the bit which said "lock-in can be split into 3 separate categories in order of badness" - I think it's pretty clear that I consider *vendor* lock-in to be much worse than *protocol* lock-in. But to deny the existance of protocol lock-in just because it isn't as bad and probably can't be fixed is extremely naive.
the fact that there are no practical alternatives for vision doesn't expose me to any avenues of exploitation.
That's a pretty good indication of why you have picked a bad analogy. In the case of network protocols, document formats and software there usually are perfectly good (often better) alternatives.
no one is in a position to potentially abuse the medium.
Again, I have to reiterate - lock-in really isn't about whether something is being abused. Just because noone is in a position to subvert something for their own ends doesn't mean that you aren't locked into it and it doesn't mean there aren't problems with being locked in.
And the fact that a lot of people use a protocol (or a file format) isn't a problem - it's what makes the protocol valuable.
In some respects it isn't a problem (makes things easier for people to exchange files, etc.). But what if that format/protocol isn't capable of doing what you need? Going back to my IPv4 vs. IPv6 example - I might choose to provide a service using IPv6 because the enhanced multicast abilities would be benficial, but I can't because all the potential users of a service are still stuck on plain old IPv4 - i.e. I'm locked in to using IPv4.
And that's also why it's so important to have these formats and protocols open and unencumbered by IPR issues, in all their various forms - because such common formats are too valuable to be entrusted to any single entity.
I'm not disagreeing with that - opening up file formats and protocols is a Good Thing and solves a lot of problems. But it doesn't solve *all* lock-in problems.
as long as you insist that the ills of lock-in stem from the lock-in itself rather than any potential for abuse, then it's hard to see how we're going to progress this
There are a lot of problems related to the abuse of lock-in, but claiming that there are no problems at all caused by unabused lock-in is just plain wrong.
Also, how do you define "abuse" - taking the Skype example, the protocol is designed to use Skype's servers and has no scope for allowing third parties to run servers. Should this be considered an abuse or just a shortsighted protocol design?
The problem I have here is that you are conflating on the one hand the problems of migrating a vary large user base with, on the other, the monopolistic abuses of power popular with a certain Seattle based software house.
Lock-in is lock-in - You don't have to be abused in order to be locked into something. Lets look at an example:
MSN - if you want to talk to all your mates who use MSN then you're going to have to use MSN too. XMPP - if you want to talk to all your mates who use XMPP then you're going to have to use XMPP too. IPv4 - If you want to access all the services on the IPv4 network then you're going to have to use IPv4 too. IPv6 - If you want to access all the services on the IPv6 network then you're going to have to use IPv6 too.
All looks pretty similar doesn't it? In each case you are locked into using a specific protocol for a specific job. The only difference in the MSN case is that, because of the protocol's design, you are also locked into using a single vendor's servers and no gatewaying between protocols can be achieved without that vendor's cooperation (yes, I'm aware of the XMPP MSN transport - it doesn't really provide real gatewaying though, you still need an MSN account).
In all these cases, various factors are locking you into using a specific protocol - this is a Bad Thing since it prevents you from picking a protocol that suits you best. However, there isn't much you can do about it. If the service you're after moves to a different protocol then you are forced to follow suit. Your friend moves from MSN to XMPP and you have to do the same if you want to continue talking to her.
Locking you into a specific _vendor_ on the other hand is something you _can_ avoid.
The fact that these stakeholder have the option not to upgrade is also atypical of the vendor lock-in scenario.
But in the long run they don't have the option not to upgrade - if major services are made available only on IPv6 then the ISPs and end users will have to upgrade or you lose access to those services.
It's not really any different from Microsoft "forcing" you to upgrade to the new version of Office. It's not as if the old version is just going to stop working - it just means you won't be able to access new content.
So my question is: are you deliberately trying to conflate the two issues? I don't think you can support the notion that they are in any way comparable.
I do see them as comparable - the requirement to upgrade to support a new protocol in order to access new content seems no different to me to the requirement to upgrade your office suite in order to access new content, or upgrade your operating system to access new content. In the latter two cases at least, the existing content won't magically stop working (in the first case that is probably not true - whilest the old protocols may be around for a long time it would seem likely that over time the services available over the old protocols would be eroded).
I think lock-in can be split into 3 separate categories in order of badness: 1. Lock in to a specific piece of software. For example, Office - if you want to exchange Office documents with people you need Office 2. Lock in to a specific service. For example, MSN, Skype, etc. - if you want to talk to your friends you are _required_ to use a single vendor's services (possibly with the option of using some 3rd party software to access the services). If they provide sucky services then that's just too bad coz you can't change vendor. 3. Lock in to a specific protocol. For example, IPv4, XMPP, SIP - if you want to talk to your friends who use XMPP then you also need to use XMPP. But it's an open protocol so you can use any vendor who provides an XMPP server (you can even run your own). If the vendor runs a crap service, you can change vendor, but you're still stuck with the same protocol - you can't choose to use SIP to talk to the XMPP network.
The first two categories are fixable by choosing to use open file formats, open protocols and decentralised protocols. The third category is probably unfixable, but I would suggest that it isn't a trivial problem.
But IPv4 addresses are a subset of IPv6, and it's likely that you'll be able to access most if not all of the current web for a long time to come.
Not entirely accurate - IPv4 addresses can be represented as IPv6 addresses, but the actual protocols are still different (even if that's hidden from the application).
it's likely that you'll be able to access most if not all of the current web for a long time to come.
When we run out of IPv4 addresses you're going to struggle to make services available through IPv4. Sure, it won't be for a few years, but it will happen. But my point is that even though no single vendor is going to force you to upgrade, you are still forced to upgrade due to the pace of progress.
Why don't you use an old 486 to browse the web? I mean, it used to render web pages just fine - noone has forced you to upgrade. However, if you want to look at the shiny new web sites your 486 probably isn't going to handle it - same thing here.
I think it's more accurate to say that TCP/IP defines the Internet, rather than that it locks you in.
I would disagree. Services on the internet use IPv4, true. However, those services can run just as well over IPv6 (for example) and the user wouldn't know that it was any different.
IPv4 has lots of problems, not least the lack of IP addresses, which forces you to use nasty hacks like NAT. However, you can't ditch IPv4 in favor of a protocol that solves these problems because you are locked into using services which are only provided over IPv4. This is why lock-in is bad - it prevents you from switching to a better technology (or at least - a technology that's better for _you_).
I'm not saying there is a solution - there probably isn't. But I'm not under any illusion - the internet locks us into using specific protocols, even if they are non-propriatory, and this has major down sides.
No one is stopping any from switching to IPv6.
That's untrue. If you run a server, you have to pick your connectivity provider very carefully if you want an IPv6 connection. Same with home users - you have to pick your ISP carefully (ignoring the fact that there are no consumer grade routers that do IPv6). Sure, there are work arounds such as tunnelling your IPv6 traffic over the IPv4 infrastructure, but this presents other problems such as increased latency, decreased bandwidth, etc.
The term, "vendor lock in" to use the full term, is where a single company controls a protocol and abuses that control to force price hikes, unnecessary upgrades and arbitrary restrictions upon its customers.
I think you've gone a bit too far with that definition. Vendor lock-in is just where a single company controls a protocol and no third parties can use it in an unrestricted way.
The company doesn't have to abuse this position - the mere fact that you _have_ to use that company's services constitutes vendor lockin, even if they are the best services in the world.
An example might be something like MSN - even though the protocol is fairly well known (through reverse engineering), and you can pick and choose your client software, the design of the MSN network *requires* you to use the central MSN servers if you want to participate in the MSN network (i.e. there is vendor lockin since you are required to use a specific vendor's services). Compare to XMPP, which is decentralised by design - you can pick and choose what server you use (and even run your own server) and still talk with people elsewhere on the XMPP network who are using different servers and clients.
But I don't think TCP/IP (the protocol that underlies the Internet) is owned by anyone as such, so it's not like we're going to get forced to pay more for a protocol "upgrade".
Well, you might be inherently forced to upgrade. When content moves to IPv6, you will need to upgrade to IPv6. The real difference here is that it isn't a single vendor forcing you to upgrade to make more money - when you need to upgrade you can pick and choose which vendor to upgrade to.
So, to summarise: please explain how can we have any meaningful lock in on the internet, and (assuming this to be possible), please also explain how this would be bad.
I think the original poster should've said "lock in" rather than "vendor lock in". There _is_ lock-in associated with the internet since that's where a lot of content is. If you want to visit Wikipedia, for example, you have to do that using HTTP over TCP over IPv4 - you can't do it using IPv6, or over Internet2 or any other network technology without some kind of gatewaying between them (probably at the application layer). Thus, you are locked in to IPv4.
This is one of the reasons why IPv6 isn't gaining ground very quickly - the server owners don't see the point in supporting IPv6 since none of the end users support it. The ISPs and end users don't see the point in supporting it since none of the server owners support it. It's a chicken-and-egg situation - why switch to IPv6 when all the content is available on IPv4 anyway?
And that's with a guaranteed fullrate, meaning that you can download at 15 Mbit, 24/7. It's like this with all the ISP's in Denmark, even the cheaper ones, which only require a little research. My own line consists of a fibre connection 25/25 Mbit, for $72 a month. Somehow they seem able to afford it, and fibre is coming up all over the country, offering faster, lower-latency, and cheaper connections. I just don't get how it is such a problem for the ISP's to deliver full flatrate, when that is what they've promised.
In the UK, at least, the problem is that the ISPs don't own the local loop. ISPs selling ADSL basically have 2 options:
1. Pay BT for bandwidth over the BT network. This involves using BT's equipment (DSLAMs, etc) and network bandwidth to transport the traffic to the ISP. 2. Use an unbundled local loop. This is only possible at certain exchanges (although more are appearing all the time) and involves the ISP putting in their own equipment at the exchange and providing their own network infrastructure to the exchanges.
Option (1) puts a fixed (fairly high) per-magabyte cost on the bandwidth, which the ISP has to pay to BT.
Option (2) is cheaper for high population areas, but isn't feasable for areas with a low population density.
I imagine that the cost of laying fibre to your door is rather too high for most ISPs - I'm not sure why it would be very cheap to do that in Denmark.
However, I would point out that $72 is rather higher than most residential ADSL connections in the UK. Normal residential connections probably average around 15ukp ($30usd), but are available at well below 10ukp ($20usd). Sure, the 10-15ukp connections are crap and very limited, but if they are sold as "unlimited" most people will buy them rather than spending the extra cash on a $72 connection. Sadly, people believe the false advertising and as far as the general public is concerned, a 10ukp "up to 8Mbps, unlimited" connection is better value for money than a 30ukp "up to 8Mbps, 50GB limit" connection.
the trouble is of course that the providers marketing departments want to advertise unlimited but the bean counters know that a certain percentage of users will use far more traffic than most and therefore will be a loss rather than a profit.
The obvious result is psuedo-unlimited services where there are no hard caps but they do everything in thier power to shaft heavy users who live in areas of high demand.
Not only do high traffic users lose out, but in order to maintain a flat-rate across all users they have to either: 1. charge stupidly high prices or 2. massively oversubscribe the network
If they do (1) then the low traffic users end up paying buckets of cash to subsidise the higher traffic users. If they do (2) then the network pretty much sucks for everyone.
The answer is pretty simple - go switch to an ISP that has a sensible business model who is honest with it's customers, rather than one that's clearly run by a moronic marketting department who believe that misleading the customer is a Good Thing.
For non-simultaneous popular media, P2P could help. Then, the provider can offload service to some connected clients.
P2P decreases the strain on the provider, but increases the strain on end-ISP networks since their users are uploading content as well as downloading it (and even worse - they are probably uploading it to other ISPs so it's not even staying within the ISP's network). Having some built in smarts in the P2P software to make it favour downloading from users topologically close to you would probably help a lot by reducing the load on the connections between autonomous networks.
When you request an unpopular item, and it's not currently available as P2P and the current maximum broadcast has been met, then you get a message that it's not available currently. Try another selection out of the million available.
That would be annoying. When I want to watch a particular video I generally don't care about the other million videos I could have - I care about the one I want to watch. I could probably live with having to wait a bit longer for the download though (although that would likely encourage people to start off multiple downloads at once so it doesn't really help - I'm still using the same total bandwidth, even if the individual downloads are going slower).
You can use multicast quite effectively for distributing like this, as long as you don't want streaming.
Only if you're dealing with popular content. If you have a single file that tousands of people are requesting at roughly the same time (i.e. maybe within a day, if you want to allow to to a day to download it) then a system as you describe works ok.
However, if you have relatively unpopular content then it doesn't work too well - to take things to the extreme, lets say you only get 1 download a day. Suddenly multicast doesn't work - you're going to have to unicast that download unless you want to force people to wait for over a day for it to download.
You may not think this is a big deal - I mean, if the content isn't popular it's not going to be using much bandwidth, right? What happens if you have thousands of unpopular files - you're going to have to unicast each one, and that'll suck your bandwidth just as much as tousands of people requesting a single unicast file.
I would be interested to know how much bandwidth cable companies such as Virgin expend with their "on demand" services. A street full of on-demand viewers fastforwarding and rewinding content is going to be pretty bandwidth-heavy (especially if they start doing HD content too).
Most consumer grade ISP services are sold as 'up to X mbps'. There is no guarantee in availability.
The issue isn't the connection speed - the problem is the total bandwidth available over a period of time.
Here in the UK, many of the smaller ISPs are selling accounts with a well publicised bandwidth limit (e.g. 30GB per month on-peak, 300GB off-peak), and making a number of different bandwidth limits available at appropriate prices. If you don't use much bandwidth then you can get a cheaper account, whilest the heavy users pay more. This is a sensible business model.
However, the larger ISPs still advertise "unlimited broadband". If you're using the word "unlimited" in your advertising then you probably can't complain when people try to max out their connection 24/7. Notably, two of the big ISPs (Tiscali and TalkTalk) have recently been complaining about the bandwidth used by people with "unlimited" accounts using the BBC's iPlayer. They sold something they couldn't provide without running at a loss on the assumption that people wouldn't use it, and now that people _are_ using what they paid for the ISPs are demanding that the BBC pay them to get them out of the hole they made for themselves.
You agreed to the fine print when you signed up for service so you really can't complain.
Most of the fine print for "unlimited" accounts just have a hand wavey "subject to fair use" clause with absolutely no indication as to what the ISP believes is "fair use". In any case, it seems like misrepresentation to me - if you advertise a product you can't then have small print that removes the very feature your adverts are using as a selling point. Advertising something as "unlimited" and then imposing limits is illegal.
You can speak loudly with your wallet, buy services from the few remaining independant ISPs
I do - I avoid buying from the ISPs who group all users together into a one-size-fits-all account. I'm not interested in a stupidly cheap service that's been overrun by the 24/7 bittorrenters and I'm not interested in a stupidly expensive service that forces me to subsidise the bittorrenters.
keep the big guys honest.
I don't hold out much hope for that. The big guys seem to be basically run by marketting departments who believe they will succeed by undercutting the competition and misleading the customer in order to do so. I don't see that this will change (hell, the cellular operators have been doing the same for years and there's no sign of them stopping any time soon) - my only hope is that the small ISPs can hold their own. The masses can stick to their massively oversubscribed AOLs whilest I use a small ISP that knows what they are doing.
Slashdot Mirror
The closest analogy therefore would be if you left your door open at home and someone - unbeknowst to the house owner - stood outside and told people they were "free to go in and take what they like". This act would not excuse anyone who then took it upon themselves to steal from your house.
That's a pretty bad analogy since you can usually tell the difference between a private residence and a public place from looking at it. This is not the case with open access points. If *all* houses looked like pubs and one had someone standing outside (unbeknowst to the owner) saying "come in for some beer", would you have any reason not to assume it was a real pub inviting people in for some beer?
Furthermore I don't think it is "fair" to simply say "well it's the responsibility of the network owner to secure their equipment, if they don't it's fair game".
The problem isn't that network owners aren't securing their access points - the problem is that it's not possible to tell the difference between an intentionally free access point and a badly configured private access point. So it seems we have two choices:
1. Make it illegal to connect to any open access point
2. Make is completely legal to connect to any open access point
There is no middle ground since it's often not possible to tell whether an AP is supposed to be open or not. I'd say that (1) would be a pretty bad state of affairs since there are a lot of legitimately open APs. It'd put the hotspot providers out of business for one, and kill all the free hotspots provided by pubs, cafes, etc.
I have a water tap in my front yard... should you assume that, without asking, you have a right to plug a hose up to it and use as much water as possible?
If you have a sprinkler in your front garden, and it's over-spraying onto my garden, should I be arrested for "stealing" your water?
It's quite clear that almost no one would want someone stealing their bandwidth for free if they knew what bandwidth was all about and the consequences, so respect that and don't use it.
That's funny - quite a few of the pubs around here intentionally set up open access points for their patrons to use. How are you suggesting I tell the difference between an intentionally open AP and an accidentally open AP?
Do not use unless you know for sure that you are allowed. An SSID of "UseThisWiFi" or similar.
It's not always clear what is public or not without explicit knowledge of what SSIDs the public access providers use.
Lets take an example. You come across an access point called "MyCloud" - there is nothing in that name that tells you it's a public access point if you didn't already know that the hotspot provider called "The Cloud" happened to exist and use "MyCloud" as their SSID.
Another example. You're at an airport and you find an open access point called "AirportWIFI", or you're at a pub and find "PubWIFI" - are you to assume that these are intentionally open for the use of the airport or pub patrons, or are they private networks that have been accidentally left open.
Another example. BT runs 802.11 hotspots, but they also provide 802.11 equipment to people subscribing to BT's ADSL connections. You find an open access point called "BT Voyager" - are you to assume that it's one of their hotspots or a private access point that's been misconfigured?
The fact is that there is no way to tell the difference between a public free hotspot and a misconfigured private access point. It seems to me that if an access point advertises itself as an open AP, and when you conntect it leases you an IP address and then gives you unrestricted access to the internet then you should be well within your rights to assume it's a public hotspot since there is no way to tell any different. On the other hand, if someone has made some attempt to secure it, even if it's just hiding the SSID, you should assume it's not intended for public use and leave it alone.
Not only that, but if your home access point is called "Linksys" then your laptop will probably quite happilly associate with any other open access point with the SSID of "Linksys" even if you don't ask it to.
dialup internet minutes via 56k modem.
:)
People still use dialup? Wow
The US has fairly clear and well enforced truth in advertising laws. Doesn't the UK?
Yep, and I don't really know why the Advertising Standards Authority doesn't come down on some of these ISPs. Although Orange has recently bee slapped down for making claims about "unlimited broadband" without mentioning a fair use policy. The ASA seems fairly happy with people saying "unlimited* broadband" and then having small print stating that fair use policy applies.
This seems wrong to me - if there is a fair use policy that places limits on how much you can download then it clearly isn't unlimited. Especially since those fair use limits actually do affect a significant number of people. It also makes it hard to compare what ISP is going to be best for you if they all hide their limits away in the small print (or worse: don't publish their limits at all).
Also, some of the fair use policies seem to place limits on much more than fair use - for example, Orange's "unlimited" GPRS tariff has a fair use policy which explicitly bans the use for instant messaging and VoIP. I wouldn't call that fair use since it won't impact other users of the system - it's just Orange preventing you from using services that would undermine their telephony business.
It seems companies can make untrue claims about their products in the UK and not get sued?
The UK culture is nowhere near as litigeous as the US. For the most part I consider that to be a Good Thing since it helps keep things like insurance costs down because there aren't so many frivolous and money-grabbing lawsuits. However, that gradually seems to be changing as more and more ambulance chasing lawyers seem to keep popping up and avertising "no win no fee" services for all sorts of situations.
False advertising claims are generally handled by the advertising standards authority, and for the most part they do a reasonable job. Although their fines are usually small and frequent enough to prevent companies from _continuing_ to misadvertise rather than big one-off fines which scare companies away from doing it in the first place. Sadly, the ASA sometimes make some bad decisions (as mentioned above), but I guess that's not a huge amount different from launching a class action and having the judge make a bad decision.
I don't _think_ there is such a thing as "class action" in the UK. And the average punter doesn't want to spend the time and money suing - unless they actually lost a lot of money they are more likley to just stop using the company in question and bad-mouth them to their friends.
Not sure if this is the case in the UK, but in the US there always seems to be the same statement--"up to X mbits". I don't know of any large ISPs that do not use that phrase.
Yup, they all use that phrase - I wasn't arguing that that was misleading (although now you come to mention it, if your network is overloaded to the point that there's no way anyone's going to get 8Mbps you probably shouldn't be allowed to advertise "up to 8Mbps").
I do not know any company that explicitly states their residential service is "unlimited".
Orange: "unlimited monthly downloads (fair use policy applies)"
Be: "Unlimited usage"
Tiscali: "a fast, affordable service with unlimited usage every month."
Toucan: "all packages are unlimited"
Pipex: "Our Pipex Max + Anytime package gives you both unlimited broadband AND unlimited local and national calls all day every day"
Thats just from a quick Google for "unlimited broadband". I hear radio ads quite frequently for "unlimited broadband" too, and infact it wasnt that long ago that Tiscali's radio ads said "Up to 8 megabyte unlimited broadband" (which was clearly a mistake and whoever let that through needs a kicking).
Right now in the US, no company has ventured out into metered billing like PlusNet has.
It's pretty common place in the UK amoungst the smaller ISPs. The bigger ones are still misleading everyone with their "unlimited" claims though.
Ultimately, many who complain a lot about residential service not being unlimited have no clue what such a service actually costs.
The customer shouldn't need to care how much the service costs the ISP. If the ISP sold them an "unlimited" then the customer has every right to expect it to be unlimited - if the ISP misadvertised it then it's the ISP's responsibility to swallow any costs required to meet their commitment.
However, the downside is that photons with higher energy than that bandgap, well, the extra energy is essentially wasted.
Since the energy that isn't converted into electricity becomes heat, would it not be sensible to mount the photovoltaic cells on the surface of thermal solar panels (i.e. water heaters)? That way, the solar panels on your roof can provide the power for your TV and the heat for your shower at the same time.
Some Linux updates come with the request for you to restart the affected service (Samba for example)
I'd prefer to restart a single service rather than the whole machine. Generally I'm not going to notice if Apache gets restarted, for example, but I am going to notice if the machine reboots since I'd lose my X session and everything that was running.
kernel updates basically leave you in the situation that modprobe has its difficulties. Being a wimp, I reboot my PC (a desktop) when that happens.
Under Fedora, RHEL and CentOS, yum installs the new kernel whilest leaving the running kernel also installed. This means you can continue running your current kernel with modprobe working and everything until you decide it's time for a reboot. So generally I only bother to reboot for security fixes.
Now, if you disagree, please explain why.
I don't really disagree - as I said, the fact that you are locked in to a specific protocol in order to access a service (which you now seem to consider is true) is probably unfixable, but that doesn't mean it doesn't cause problems.
The original poster said: By 2020, the current internet will have a level of lock-in that makes Windows look disposable. "Faster" and "safer" will have a tough time overcoming "empty".
And I firmly believe he's right (infact, I think that this level of lock-in happened long ago). So much content and services are available over the internet (or more specifically: IPv4) that you are pretty much locked in to using IPv4.
If someone comes out with a new protocol (which has already happened with IPv6) or even a whole new infrastructure, that new technology is going to have real trouble gaining ground for the simple reason that it has no content/services on it - the content/services are still all on IPv4 (or available through both). The mere fact that IPv4 is "good enough" for most uses means that there is little incentive for everyone to upgrade.
Remember that to roll out an "internet replacement" you're going to need to get the server owners, the connectivity providers and the end users to all upgrade. If one of them doesn't feel it's worth their while then noone gets to play with the "internet replacement" even if they want to.
Sure, you can roll out IPv6 on your own network (I did years ago) and run IPv6 and IPv4 in parallel, but the vast majority of your traffic is going to be IPv4 for years to come unless a major reason comes up for the server owners and ISPs to expend the money needed to upgrade.
With IPv6, there are major reasons for upgrading:
- Do away with NAT. This is mostly going to affect the end users and server owners since P2P protocols such as VoIP are becoming more popular. Unfortunately, P2P does work through NAT (but is less reliable) so much of the push to upgrade is gone there and then.
- IP address shortage. This is possibly the bigger reason. Even with NAT we _are_ going to run out of IPv4 addresses at some point in the future. However, I feel this is going to be very much like the Y2K problem - everyone can see it coming but noone is willing to spend the money to plan for the future. Eventually this problem is going to hit us hard and there will be an "oh shit the sky is falling" moment where everyone will end up spending way more money upgrading than they would've done if they had upgraded earlier.
No one who has lived in my house in 60 years has ever picked up the phone and it not worked.
That is a different experience than those who use this service have.
I'm the last person to defend Skype, but it seems your analysis is flawed. You have compared the experiences of a single household with the experiences of several thousand slashdotters.
Personally, I have had outages on phone lines on several occasions. I can think of at least 2 occasions when my BT line has completely failed and probably around 3 occsions when my NTL line has failed.
I'm not saying VoIP is less reliable than the PSTN (infact I'm sure it is), but I am saying that the basis for your "proof" is completely bogus.
Oh, please! How is that to be considered a lock in? Does my use of SIP prevent me from using XMPP? No it does not.
How does your use of Office prevent you from using Open Office? It doesn't.
How does your use of MSN prevent you from using XMPP? It doesn't.
Yet, both MS Office and MSN are considered cases of vendor lock-in _because_ you are tied to using a specific program/service/protocol in order to access specific content.
Will the XMPP corporation change the protocol so that I'm required to buy their software all over again? No, that will not happen.
XMPP corporation? XMPP is the IETF standard instant messaging protocol - there is no XMPP corporation.
And yes, the XMPP protocol does change and you are required to upgrade your XMPP client if you want to use the new functionality. Just as Office documents change to incorporate new functionality and you are required to upgrade your copy of Office in order to use it.
In the case of IPv4 and IPv6, that doesn't apply. I can migrate any time I want to. The fact that there's not much point until most of the rest of the world does likewise doesn't mean I'm locked in
You can migrate from MS Office at any time you want. The alternatives won't be much use (assuming they can't use Office documents) if you want to exchange documents with the rest of the world, but there's nothing stopping you switching. I really don't see a difference:
* You can change from IPv4 to IPv6 but there's no point since the content is only available over IPv4.
* You can change from Office to $other_office_suite but there's no point since the content is only available in Office format.
* You can change from MSN to XMPP but there's no point since the people you want to talk to are only on MSN.
All looks pretty similar to me.
I did, and I chose not to address your point because I felt (and still feel) that you were proceeding from a false premise.
Not addressing my point is one thing. However, what you did was ignore my point and willfully assume I believed the exact opposite of what I stated I believed.
[Office] In particular, there are also issues of deliberate breaking of backwards compatibility to force othewise undeeded upgrades
IPv6 breaks backwards compatibility with IPv4 in the same way. Compare:
* You want to open an Office document that was saved from a newer version of Office, thus you have to upgrade your version of Office too. But all your old documents still work just fine with your existing Office.
* You want to use a service that is only available over IPv6, thus you have to upgrade your IP stack. But all the old services still work fine over IPv4*
In both cases, you are not _required_ to upgrade, however you do need to upgrade *IFF* you want to access new stuff.
(* as mentioned, this isn't necessarilly the case - if IPv6 takes off then IPv4 services will be erroded and replaced with IPv6-only services. In a way, it makes it _worse_ than Office since you can nolonger guarantee your existing stuff will continue working without an upgrade).
This I do not accept as lock in. There is no key holder here and no potential for abuse
As I have stated before, I do not believe the potential for abuse is a defining feature of lock-in. Sure, if you are locked in then there may be an increased potential for abuse, but that doesn't mean that all lock-in scenarios can be abused.
So I still think you're attempting to trivialise a serious industry problem by conflating it with a completely unrelated problem on the one hand
I'm not intending to trivialise it at all. There _is_ a serious industry problem with lock-in, and that is a big problem. But to believe that there is no such thing as lock-in (and no disadvantages) in situations where a single organisation isn't in a position to abuse people seems extremely naive.
Rather than trying to trivialise the "vendor lock-in" problem, I am trying t
correct... if no firewall is present... if a firewall block UDP packet, then relying is a must...
Or the user could get a clue and poke the correct holes in the firewall. I'm not convinced that engineering protocols to try and bypass the user's own firewall is a bright idea - firewalls are there for a reason.
EVERY other software distribution has [frequent, but not necessarily monthly] updates that require a restart like this.
Umm.. no... My linux machines very rarely need to be rebooted in order to receive security updates - the security updates quietly happen nightly and I rarely need to bother about them. The Windows machines, on the other hand, seem to need a reboot every month.
Basically the problem can be put down to the combination of 2 problems:
1. A operating system with a flawed update system that requires regular reboots for all users at roughly the same time (Windows Update).
2. A flawed protocol that requires all users access the same cluster of centralised servers (Skype).
The solution: avoid using such flawed systems - there are better alternatives available.
And this would be a perfect example of why centralised networks are a Stupid Idea.
Use a decentralised technology such as SIP - if something catestrophic happens, only a few users are affected rather than taking out the whole network and if your service provider provides a sucky quality of service you can go move to a different one (infact you don't even need a server if you're not having to negotiate things like NAT - SIP runs quite happilly as a peer-to-peer application).
You provide a false choice by suggesting ISPs could possibly choose to sell service without any oversubsription. Say the 10mbit service my company offers were sold assuming 100% usage 100% all the time. Depending on infrastructure costs and typical bandwidth costs for a decent size ISP in the US, you would be talking around $500 to $1000 per month. Say you have 1,000,000 customers in a modest size city. Are you suggesting that the ISP should plan on delivering 10 Petabits of data to that community? I'm not sure that that kind of bandwidth is even feasible with current technologies.
If they want to sell "unlimited" connections then yes, they need to provide enough bandwidth to support what they are selling and set their prices appropriately to cover the cost.
Alternatively a more sensible suggestion would be to stop misadvertising their services as "unlimited" and instead include information as to what the limits actually are. Also, providing several different products with different limits and appropriate pricing seems a good idea to keep both the light users and the heavier users happy, rather than having one group of users subsidising the other.
Really, if you advertise a service and use the fact that it's "unlimited" as a selling point, you damned well can't complain when people try to use it in an unlimited way.
An ISP could just change their terms of service and next month the customer can decide if they want to stay.
Herein lies the problem - when everyone else in the market is advertising "unlimited" services, your service doesn't _look_ attactive if it's similarly priced but has published limits even though the "unlimited" services probably have similar limits which just aren't published.
Changing your terms of service also leads to bad publicity (those few unprofitable heavy users can make a lot of noise - I saw it happen when PlusNet changed their terms in this way). And even the less well informed people who don't overuse the service may feel hard done by since you appear to be taking something away that they paid for (even though they aren't using it and their service will probably actually _improve_ as a result).
So in a market where everyone else is misrepresenting their products it's pretty hard to be truthful whilest still looking competetive.
Thankfully here in the UK, many of the smaller ISPs have stopped advertising their connections as "unlimited" and in a number of cases have put together a good explanation of why no ISP can offer unlimited internet connections. I would be interested in how many potential customers ignore that and still believe an "unlimited" connection is better value for money so go elsewhere though.
Not only is the problem not inconveniencing anyone sane
Completely untrue.
If you're using SIP as your instant messaging protocol and I'm using XMPP, one of us has to change protocol (or use both) - that is inconvenient.
Similarly, I have some servers that are only accessible over IPv6 (because I don't have enough public IPv4 addresses to go around). These are specifically servers that I only need to access from locations that I know have IPv6 connectivity so it isn't a problem, but it does mean that if I needed someone with only IPv4 access to connect to them I'd have to jump through some hoops - that is inconvenient.
By your argument a lock-in is a lock-in, and therefore that's a Bad Thing, or at least as bad as any other lock in.
That is explicitly _not_ what I said. If you had read my post you would've seen the bit which said "lock-in can be split into 3 separate categories in order of badness" - I think it's pretty clear that I consider *vendor* lock-in to be much worse than *protocol* lock-in. But to deny the existance of protocol lock-in just because it isn't as bad and probably can't be fixed is extremely naive.
the fact that there are no practical alternatives for vision doesn't expose me to any avenues of exploitation.
That's a pretty good indication of why you have picked a bad analogy. In the case of network protocols, document formats and software there usually are perfectly good (often better) alternatives.
no one is in a position to potentially abuse the medium.
Again, I have to reiterate - lock-in really isn't about whether something is being abused. Just because noone is in a position to subvert something for their own ends doesn't mean that you aren't locked into it and it doesn't mean there aren't problems with being locked in.
And the fact that a lot of people use a protocol (or a file format) isn't a problem - it's what makes the protocol valuable.
In some respects it isn't a problem (makes things easier for people to exchange files, etc.). But what if that format/protocol isn't capable of doing what you need? Going back to my IPv4 vs. IPv6 example - I might choose to provide a service using IPv6 because the enhanced multicast abilities would be benficial, but I can't because all the potential users of a service are still stuck on plain old IPv4 - i.e. I'm locked in to using IPv4.
And that's also why it's so important to have these formats and protocols open and unencumbered by IPR issues, in all their various forms - because such common formats are too valuable to be entrusted to any single entity.
I'm not disagreeing with that - opening up file formats and protocols is a Good Thing and solves a lot of problems. But it doesn't solve *all* lock-in problems.
as long as you insist that the ills of lock-in stem from the lock-in itself rather than any potential for abuse, then it's hard to see how we're going to progress this
There are a lot of problems related to the abuse of lock-in, but claiming that there are no problems at all caused by unabused lock-in is just plain wrong.
Also, how do you define "abuse" - taking the Skype example, the protocol is designed to use Skype's servers and has no scope for allowing third parties to run servers. Should this be considered an abuse or just a shortsighted protocol design?
The problem I have here is that you are conflating on the one hand the problems of migrating a vary large user base with, on the other, the monopolistic abuses of power popular with a certain Seattle based software house.
Lock-in is lock-in - You don't have to be abused in order to be locked into something. Lets look at an example:
MSN - if you want to talk to all your mates who use MSN then you're going to have to use MSN too.
XMPP - if you want to talk to all your mates who use XMPP then you're going to have to use XMPP too.
IPv4 - If you want to access all the services on the IPv4 network then you're going to have to use IPv4 too.
IPv6 - If you want to access all the services on the IPv6 network then you're going to have to use IPv6 too.
All looks pretty similar doesn't it? In each case you are locked into using a specific protocol for a specific job. The only difference in the MSN case is that, because of the protocol's design, you are also locked into using a single vendor's servers and no gatewaying between protocols can be achieved without that vendor's cooperation (yes, I'm aware of the XMPP MSN transport - it doesn't really provide real gatewaying though, you still need an MSN account).
In all these cases, various factors are locking you into using a specific protocol - this is a Bad Thing since it prevents you from picking a protocol that suits you best. However, there isn't much you can do about it. If the service you're after moves to a different protocol then you are forced to follow suit. Your friend moves from MSN to XMPP and you have to do the same if you want to continue talking to her.
Locking you into a specific _vendor_ on the other hand is something you _can_ avoid.
The fact that these stakeholder have the option not to upgrade is also atypical of the vendor lock-in scenario.
But in the long run they don't have the option not to upgrade - if major services are made available only on IPv6 then the ISPs and end users will have to upgrade or you lose access to those services.
It's not really any different from Microsoft "forcing" you to upgrade to the new version of Office. It's not as if the old version is just going to stop working - it just means you won't be able to access new content.
So my question is: are you deliberately trying to conflate the two issues? I don't think you can support the notion that they are in any way comparable.
I do see them as comparable - the requirement to upgrade to support a new protocol in order to access new content seems no different to me to the requirement to upgrade your office suite in order to access new content, or upgrade your operating system to access new content. In the latter two cases at least, the existing content won't magically stop working (in the first case that is probably not true - whilest the old protocols may be around for a long time it would seem likely that over time the services available over the old protocols would be eroded).
I think lock-in can be split into 3 separate categories in order of badness:
1. Lock in to a specific piece of software. For example, Office - if you want to exchange Office documents with people you need Office
2. Lock in to a specific service. For example, MSN, Skype, etc. - if you want to talk to your friends you are _required_ to use a single vendor's services (possibly with the option of using some 3rd party software to access the services). If they provide sucky services then that's just too bad coz you can't change vendor.
3. Lock in to a specific protocol. For example, IPv4, XMPP, SIP - if you want to talk to your friends who use XMPP then you also need to use XMPP. But it's an open protocol so you can use any vendor who provides an XMPP server (you can even run your own). If the vendor runs a crap service, you can change vendor, but you're still stuck with the same protocol - you can't choose to use SIP to talk to the XMPP network.
The first two categories are fixable by choosing to use open file formats, open protocols and decentralised protocols. The third category is probably unfixable, but I would suggest that it isn't a trivial problem.
But IPv4 addresses are a subset of IPv6, and it's likely that you'll be able to access most if not all of the current web for a long time to come.
Not entirely accurate - IPv4 addresses can be represented as IPv6 addresses, but the actual protocols are still different (even if that's hidden from the application).
it's likely that you'll be able to access most if not all of the current web for a long time to come.
When we run out of IPv4 addresses you're going to struggle to make services available through IPv4. Sure, it won't be for a few years, but it will happen. But my point is that even though no single vendor is going to force you to upgrade, you are still forced to upgrade due to the pace of progress.
Why don't you use an old 486 to browse the web? I mean, it used to render web pages just fine - noone has forced you to upgrade. However, if you want to look at the shiny new web sites your 486 probably isn't going to handle it - same thing here.
I think it's more accurate to say that TCP/IP defines the Internet, rather than that it locks you in.
I would disagree. Services on the internet use IPv4, true. However, those services can run just as well over IPv6 (for example) and the user wouldn't know that it was any different.
IPv4 has lots of problems, not least the lack of IP addresses, which forces you to use nasty hacks like NAT. However, you can't ditch IPv4 in favor of a protocol that solves these problems because you are locked into using services which are only provided over IPv4. This is why lock-in is bad - it prevents you from switching to a better technology (or at least - a technology that's better for _you_).
I'm not saying there is a solution - there probably isn't. But I'm not under any illusion - the internet locks us into using specific protocols, even if they are non-propriatory, and this has major down sides.
No one is stopping any from switching to IPv6.
That's untrue. If you run a server, you have to pick your connectivity provider very carefully if you want an IPv6 connection. Same with home users - you have to pick your ISP carefully (ignoring the fact that there are no consumer grade routers that do IPv6). Sure, there are work arounds such as tunnelling your IPv6 traffic over the IPv4 infrastructure, but this presents other problems such as increased latency, decreased bandwidth, etc.
The term, "vendor lock in" to use the full term, is where a single company controls a protocol and abuses that control to force price hikes, unnecessary upgrades and arbitrary restrictions upon its customers.
I think you've gone a bit too far with that definition. Vendor lock-in is just where a single company controls a protocol and no third parties can use it in an unrestricted way.
The company doesn't have to abuse this position - the mere fact that you _have_ to use that company's services constitutes vendor lockin, even if they are the best services in the world.
An example might be something like MSN - even though the protocol is fairly well known (through reverse engineering), and you can pick and choose your client software, the design of the MSN network *requires* you to use the central MSN servers if you want to participate in the MSN network (i.e. there is vendor lockin since you are required to use a specific vendor's services). Compare to XMPP, which is decentralised by design - you can pick and choose what server you use (and even run your own server) and still talk with people elsewhere on the XMPP network who are using different servers and clients.
But I don't think TCP/IP (the protocol that underlies the Internet) is owned by anyone as such, so it's not like we're going to get forced to pay more for a protocol "upgrade".
Well, you might be inherently forced to upgrade. When content moves to IPv6, you will need to upgrade to IPv6. The real difference here is that it isn't a single vendor forcing you to upgrade to make more money - when you need to upgrade you can pick and choose which vendor to upgrade to.
So, to summarise: please explain how can we have any meaningful lock in on the internet, and (assuming this to be possible), please also explain how this would be bad.
I think the original poster should've said "lock in" rather than "vendor lock in". There _is_ lock-in associated with the internet since that's where a lot of content is. If you want to visit Wikipedia, for example, you have to do that using HTTP over TCP over IPv4 - you can't do it using IPv6, or over Internet2 or any other network technology without some kind of gatewaying between them (probably at the application layer). Thus, you are locked in to IPv4.
This is one of the reasons why IPv6 isn't gaining ground very quickly - the server owners don't see the point in supporting IPv6 since none of the end users support it. The ISPs and end users don't see the point in supporting it since none of the server owners support it. It's a chicken-and-egg situation - why switch to IPv6 when all the content is available on IPv4 anyway?
And that's with a guaranteed fullrate, meaning that you can download at 15 Mbit, 24/7. It's like this with all the ISP's in Denmark, even the cheaper ones, which only require a little research. My own line consists of a fibre connection 25/25 Mbit, for $72 a month. Somehow they seem able to afford it, and fibre is coming up all over the country, offering faster, lower-latency, and cheaper connections. I just don't get how it is such a problem for the ISP's to deliver full flatrate, when that is what they've promised.
In the UK, at least, the problem is that the ISPs don't own the local loop. ISPs selling ADSL basically have 2 options:
1. Pay BT for bandwidth over the BT network. This involves using BT's equipment (DSLAMs, etc) and network bandwidth to transport the traffic to the ISP.
2. Use an unbundled local loop. This is only possible at certain exchanges (although more are appearing all the time) and involves the ISP putting in their own equipment at the exchange and providing their own network infrastructure to the exchanges.
Option (1) puts a fixed (fairly high) per-magabyte cost on the bandwidth, which the ISP has to pay to BT.
Option (2) is cheaper for high population areas, but isn't feasable for areas with a low population density.
I imagine that the cost of laying fibre to your door is rather too high for most ISPs - I'm not sure why it would be very cheap to do that in Denmark.
However, I would point out that $72 is rather higher than most residential ADSL connections in the UK. Normal residential connections probably average around 15ukp ($30usd), but are available at well below 10ukp ($20usd). Sure, the 10-15ukp connections are crap and very limited, but if they are sold as "unlimited" most people will buy them rather than spending the extra cash on a $72 connection. Sadly, people believe the false advertising and as far as the general public is concerned, a 10ukp "up to 8Mbps, unlimited" connection is better value for money than a 30ukp "up to 8Mbps, 50GB limit" connection.
the trouble is of course that the providers marketing departments want to advertise unlimited but the bean counters know that a certain percentage of users will use far more traffic than most and therefore will be a loss rather than a profit.
The obvious result is psuedo-unlimited services where there are no hard caps but they do everything in thier power to shaft heavy users who live in areas of high demand.
Not only do high traffic users lose out, but in order to maintain a flat-rate across all users they have to either:
1. charge stupidly high prices
or
2. massively oversubscribe the network
If they do (1) then the low traffic users end up paying buckets of cash to subsidise the higher traffic users. If they do (2) then the network pretty much sucks for everyone.
The answer is pretty simple - go switch to an ISP that has a sensible business model who is honest with it's customers, rather than one that's clearly run by a moronic marketting department who believe that misleading the customer is a Good Thing.
For non-simultaneous popular media, P2P could help. Then, the provider can offload service to some connected clients.
P2P decreases the strain on the provider, but increases the strain on end-ISP networks since their users are uploading content as well as downloading it (and even worse - they are probably uploading it to other ISPs so it's not even staying within the ISP's network). Having some built in smarts in the P2P software to make it favour downloading from users topologically close to you would probably help a lot by reducing the load on the connections between autonomous networks.
When you request an unpopular item, and it's not currently available as P2P and the current maximum broadcast has been met, then you get a message that it's not available currently. Try another selection out of the million available.
That would be annoying. When I want to watch a particular video I generally don't care about the other million videos I could have - I care about the one I want to watch. I could probably live with having to wait a bit longer for the download though (although that would likely encourage people to start off multiple downloads at once so it doesn't really help - I'm still using the same total bandwidth, even if the individual downloads are going slower).
You can use multicast quite effectively for distributing like this, as long as you don't want streaming.
Only if you're dealing with popular content. If you have a single file that tousands of people are requesting at roughly the same time (i.e. maybe within a day, if you want to allow to to a day to download it) then a system as you describe works ok.
However, if you have relatively unpopular content then it doesn't work too well - to take things to the extreme, lets say you only get 1 download a day. Suddenly multicast doesn't work - you're going to have to unicast that download unless you want to force people to wait for over a day for it to download.
You may not think this is a big deal - I mean, if the content isn't popular it's not going to be using much bandwidth, right? What happens if you have thousands of unpopular files - you're going to have to unicast each one, and that'll suck your bandwidth just as much as tousands of people requesting a single unicast file.
I would be interested to know how much bandwidth cable companies such as Virgin expend with their "on demand" services. A street full of on-demand viewers fastforwarding and rewinding content is going to be pretty bandwidth-heavy (especially if they start doing HD content too).
Most consumer grade ISP services are sold as 'up to X mbps'. There is no guarantee in availability.
The issue isn't the connection speed - the problem is the total bandwidth available over a period of time.
Here in the UK, many of the smaller ISPs are selling accounts with a well publicised bandwidth limit (e.g. 30GB per month on-peak, 300GB off-peak), and making a number of different bandwidth limits available at appropriate prices. If you don't use much bandwidth then you can get a cheaper account, whilest the heavy users pay more. This is a sensible business model.
However, the larger ISPs still advertise "unlimited broadband". If you're using the word "unlimited" in your advertising then you probably can't complain when people try to max out their connection 24/7. Notably, two of the big ISPs (Tiscali and TalkTalk) have recently been complaining about the bandwidth used by people with "unlimited" accounts using the BBC's iPlayer. They sold something they couldn't provide without running at a loss on the assumption that people wouldn't use it, and now that people _are_ using what they paid for the ISPs are demanding that the BBC pay them to get them out of the hole they made for themselves.
You agreed to the fine print when you signed up for service so you really can't complain.
Most of the fine print for "unlimited" accounts just have a hand wavey "subject to fair use" clause with absolutely no indication as to what the ISP believes is "fair use". In any case, it seems like misrepresentation to me - if you advertise a product you can't then have small print that removes the very feature your adverts are using as a selling point. Advertising something as "unlimited" and then imposing limits is illegal.
You can speak loudly with your wallet, buy services from the few remaining independant ISPs
I do - I avoid buying from the ISPs who group all users together into a one-size-fits-all account. I'm not interested in a stupidly cheap service that's been overrun by the 24/7 bittorrenters and I'm not interested in a stupidly expensive service that forces me to subsidise the bittorrenters.
keep the big guys honest.
I don't hold out much hope for that. The big guys seem to be basically run by marketting departments who believe they will succeed by undercutting the competition and misleading the customer in order to do so. I don't see that this will change (hell, the cellular operators have been doing the same for years and there's no sign of them stopping any time soon) - my only hope is that the small ISPs can hold their own. The masses can stick to their massively oversubscribed AOLs whilest I use a small ISP that knows what they are doing.
Oops:
:)
there is centralised service required
Clearly I meant *NO* centralised service required