NO the biggest 32-bit integer is 2^32-1 the *number of values* is 2^32 the first value is 0 but even that is not the whole picture, because for instance all IPs that start with 192.168 are reserved for private use and cannot be given out nor should be used outside a LAN and there are other reserved numbers another aspect of the problem is that addresses are given out in blocks which are hardly ever completely used the same problem will possibly confront the world's telephone systems someday but no worries, the IT industry is, as was said, in the process of switching to 128-bit IPs 128 bits is enough to give every electron in our galaxy its own address assuming we dont waste too many addresses
OMG. Is this what physicists feel like when they read the comments under physics vidoes ? Back in the days before the web I would often remember the straight IP address instead of a domain name. Only for places I used regularly though. These days I don't even know what my own IP address is.
Allan Richardson No I am not paying extra. My ISP was the first one ever in the UK and they were pretty innovative developing new technologies which most other later ISP's adopted. They were true techies. They offered an excellent service without relying on irrelevant bells and whistles, or questionable business practices such as AOL. Unfortunately they were eventually bought by a larger ISP, and once the suits took over, the quality of the service started to drop. Tech support used to be by Brits who really knew what they were talking about. Now it's been outsourced to India where you get some random person with a hard to understand accent reading through a script.
Technically the 32 bit scheme is infinitely extensible by using NAT. Also, the video acts like the devices in your house take up extra adresses, but that is not true. They will be off of your single home IP, and internally addressed with at 10.x or 192.x address
Not infinitely, NAT extends IP addresses by combining IP address and port number to uniquely identify a computer. Port numbers are limited to 65,535 (16 bits). So NAT can extend 32bit addressing to at most 48bit, usually much less because normal work on a computer (such as watching RUclips) needs a dozen or two of port numbers.
Was the problem that Internet Servers needed unique IP addressing? Most Net users are normally clients and dont need the full address exclusivity, just a unique, dynamic or translated return path. Was IPv5 going to be a 64 bit system, but the technology jump to 128 came in faster than IPv5 could be commercially developed
nimim. Marko Mikkilä I do believe that the Defense Department handed over full control of the Internet to five regional consortia of universities in the U.S. quite some time ago, now. It is they that own the backbone, in any case, and I don't believe that the Defense Department is paying for any of it at this point. So Professor Clewett can relax! As for his wondering about domain names, they were instituted by the same team that developed IP and have always been part of the Internet.
Steven Titus For website hosting on a dedicated address without depending on a centralized proxy (which NAT based upon), you need to buy the right to a IPv6 address from the IP address managing group (name was in the video, but I don't remember off the top of my head), or you need to find a willing seller of a IPv4 address holder. Market Pricing via supply and demand is a temporary and decentralized solution to this as you, but if the expansion of the Internet of Things keeps up its exponential pace, then the demand will overwhelm the supply of sellers to the point where prices of IPv4 addresses will skyrocket, thus forcing individuals and companies to either accept the pyramid like structure and centralization that NAT forms (which we somewhat have now), and/or buy the rights to never used IPv6 addresses. Big ISPs like comcast already use NAT for their customers. All of the comcast customers in my area use the same public IPv4 address from the regional hub, so then from the centralized hub, every customer is given a unique IPv4 address from the viewpoint of the central hub via NAT as an intranet. All router-based LAN networks work the same way - all devices have their own IPv4 (and usually IPv6 address as well) from the view point of the router as a intranet, and connect to the rest of the internet via NAT on your router and your modem.
Kyle Herbig Yes, I know how NAT works. I was just making an observation that he didn't mention it in the video. At 2:16 he could have mentioned NAT. Its just a little bit misleading, that's all i'm saying.
+Steven Titus Yea, I too thought it was a bit strange to leave that out. Since the statement that all your devices need a unique address just isn't true.. I even heard from people in India that they actually got internet via some sort of LAN network with their ISP. Which, although not a great solution, is a solution to the problem of running out of addresses. (Btw I understand that in the end we just need ipv6 )
+Кэтишкэт Старье 7 number addresses + a point of origin. And if i counted correctly there are about 40 symbols on each stargate - 1 for the point of origin. Without all the working its about 15.4 million 15,380,937. Pegasus ( at least in real life ) is a dwarf galaxy and it only has a few billion stars, so lets be generous and say 10 billion. Using the parts of the drake equation we need it comes out at about 3.2 million planets that can support human like life. So no there are plenty.
Isaac Karjala To be honest, about a decade ago I would have agreed with you. There was a point in time where I really wanted to have every computer on my home network with a routable address for protocols like you mentioned. But the way that so many Internet-connected home devices these days work, you go through some third-party portal site that interacts with the device using a proprietary protocol. My Nest thermostats (yeah, I have two in my house for different zones), for instance, are stuck behind NAT but the Nest website straightens all that out for me. Western Digital has a sort of similar service for their network attached storage devices. It's like the future is to hide protocol details behind some cloud service. I hate the phrase "Internet 2.0", but that really seems appropriate ;)
What? I was typing a comment and RUclips just went to the next video? I was saying there are fewer than 4,294,967,296 IP addresses in use on the Internet. All the RFC 1918 addresses, as well as numbers beginning with 0 ("Discovery"), 127 ("Loopback") and 224-255 (Multicast and "Reserved") make up a big chunk of addresses you'll never see, along with a bunch of others. My calculation that is 592,708,864 addresses (13.8% of the number given) not usable. This doesn't even include address ranges like 1.1.1.x and 1.2.3.x that are informally reserved, plus some companies and the DOD have large blocks that may not actually be in use.
its actually more complicated than that most homes get one ipv4 address and each device gets a non routable address behind that address much like how apartment numbers work
This is supposed to change with IPv6 though, isn't it? With the incredible amount of available addresses, basically every device in a household should be able to get a public address themselves.
no less or more secure. saying its non routable is not 100% true. My understanding is the outer header for the packet contains the ipv4 address for your house, then your local address is at the next stage of the header. you can start pinging for the local adresses, guessing at local routing names (usually people will have 192.168.0.1 or some other std format), and find the full address for a computer behind a local network. Also most attacks originate from the computer itself in the form of malware, in that case you already know the address. also I don't know the ratio, but most budget internet connections do not assign a unique ip address to each home, instead sharing them by assigning a dynamic IP address each time you reset your connection (and sometimes while its still running). usually doesn't cause an issue, but you can't run a home server accessible offsite, or be the server for an online game, or home host a website with a dynamic ip (easily). switching to ipv6 would make things alot easier on many levels, and in reality the added layer of obscurity in ipv4 is only an annoyance to an attacker, not something that significantly reduces the occurence of attacks.
It is actually the 4,294,967,286th in a series that starts at zero , but by convention we start counting at 1, not zero. So it rolls over to the next place even though the next register doesn't exist in the computer. It gets to 4,294,967,296 because we talk about it as if we started counting at 1.. Another example is when a dollar equals 100 cents - where this takes three places. But you can count to 100 if you start at zero as the first, and 99 as the hundredth. That only requires two places. That's the difference between computers, where IP tables are maintained, and everyday reckoning. And yes, iIt can be confusing.
It's not confusing if you ask proper question. How many 1-digit numbers can you write? 10 - first 2-digit number. How many 2-digit numbers can you write? 100 - first 3-digit number. ... How many 32-digit numbers can you write? 1 0000 0000 0000 0000 0000 0000 0000 0000 - first 33-digit number. And it's true for any system as long as 0 mean zero and 1 mean one :)
Carpey computers count from 0, and we count from 1, e.g. the first number would be 0000000 in 8 bit, if you count from 0 it would be 11111111111111111111111111111111
That's not what he said... They govern the IP tables... The origins of the internet is TCP/IP (the protocol that every computer is using to talk to each other) and that was created in CERN in Switzerland.
OverSoft Well, the technology of the internet was created by the United States Defense Department(or at the very least was funded by it), but the WWW was created at CERN.
TheGreatRakatan They're not the creators of the internet, they just determine public ip adresses. The first people that connected computers to communicate with each other and they've called it net were from some California university. I'm sorry, if I'm wrong, but this is what i remember from the CNA (Cisco Networking Academy) interpretation.
You are correct if you start to count at one. His main point was that an 8-bit number has 256 different values, i.e., 0 to 255. If you encode binary numbers in two's complement, then an 8-bit number would range from -128 to 127 (so can only count to 127 in this case), but still have 256 different values.
+Byamarro2 Well internal (or private) IP's are a HUGE contributor to conserving IP addresses, as essentially every household need only 1 external IP address (companies might have a few, as they might have a few different connections outbound but still, likely less than 10 for hundreds if not thousands of machines)
Brady, you asked about content for the new computing channel: James has covered key elements of domain and ip fundamentals really well there. Maybe some more of this type of content?
Why they decide to use 128 bit instead of 64? 64 bits it already 18 quintillions of addresses. But to fill 128bit you need to convert all mass of the earth into devices, and each one should weight 5*10^(-11) gram! That is ridiculous!
They probably want to avoid the same problem in the future. Humans are notoriously bad at predicting the future, people said the car would never replace horses, no one would ever need a computer in their house, people would never need more than a megabyte of memory etc. Back in the 80s and 90s they probably figured that there would never be billions of computers using IP addresses or that by that time computers would have advanced past the severe processing limitations they had then. Given the proliferation of computer technology who knows how many devices will require IP addresses in 50 years or 100 years etc? Better to just make it 128-bit and never worry about it again, its not like its too demanding in processing power. There are also more technical and security reasons as well
whos to say how the future will work like thereine said who knows what will use them perhaps nano machine will use them up or small drones best go with the big number
But that's only planet earth! So small! What about the entire galaxy? Or bigger? Sure we may not get there, but if we do, it'll be well worth the investment. Also, imagine fixing 18 quintillion devices to support IPv24658. Doesn't sound very easy, does it?
What most forget is that a single device might use multiple addresses at the same time. There are multicast or unicast addresses which devices can "use" to actually hook into a stream. That's already happening with IPv4, though there are just a few multicast addresses out there ^^ There are a lot of special addresses which are locked for special uses. For example the IPv4 reserves a whole class A subnet just for the local loopback address. This address can only be used locally on each device and always refers to the device itself. The whole 127.X.X.X space is wasted for that ^^. That's 16 million addresses or 1/256.of all addresses in the IPv4 address space. It would also simplify the assignment of IP addresses. You could actually define certain addresses to be bound to the location on earth. So each square meter might have a million addresses available. A seperate address space might be used for mobile devices (as they very likely change position). The GUID also uses 128 bits. It's "said" that the possibility that the same GUID is picked twice is very very unlikely. Same for MD5 hashes, they are also 128bits long. Since modern computer hardware is already able to natively handle 128bit values it's no suprise they picked 128bits for IPv6.
Network Address Translation The local network they run on shares the same IP to the router, and then the router redirects the traffic to the appropriate internal address. If you want a more accurate description we have wikipedia for that.
If people are still wondering. Domain Name is basically the 'mask' of its IP. Then a DNS is used to translate the Domain Name to its IP Address, in other words 'see what's behind the mask'. Think of it as a Phonebook. Doman Name is the Name. IP Address is the Number. DNS (Domain Name Server) is the Phonebook itself. Very brief and vague explanation but I tried to keep it simple.
when they say, "computer can only count up to 4 billion" they mean when it specifically comes to this address thing. the thing is, those old computers can only handle numbers up to 4 billion at once, which it needs to do for addresses. but when it comes to stuff at the user level, such as a calculator or storing the number of youtube views, it can take the number in chunks, and have no problem handling larger numbers.
Apparently RUclips was forced to update their view count system to 64 bit since Gangnam style got over 2,147,483,647 views, which is the highest possible number in a signed 32 bit system. Maybe they should have went with unsigned integers to hold up to 4,294,967,295 views. It's not like there's negative views.
ComandanteJ The only difference is there is no classful boundary or broadcast addresses and hex characters versus decimal. I guess the difficulty of the process is debatable. I found it to be about the same since your main concern in both ipv4 and ipv6 is how many bits represent the subnet bits and how many represent host bits. In IPv6 you are converting to binary from hex instead of decimal. In my opinion the difficulty of the process is the same.
This guy's talking as if every device gets a different public IP. Routers give local IPs and then the modem is the only device getting a public IP.Please slap that fool.
Including public/private address spaces and mechanisms like NAT in this video would just add confusion on the topic. It also doesn't help since the problem described still applies.
He could've mentioned that in a regular home, it really doesn't matter how many devices you use, because your router is probably set up as a NAT, meaning it counts as 1 IP globally.
NAT is still used by almost everyone. The good thing is that it's 99% automatic, so no one notices it. Any time you connect your computer to a wifi hotspot, or a router at home or work, that router gives it an address in the 192.168.x.x block. The router then uses NAT to let you talk to the outside world. This means the ISPs only have to give one unique IP address to each customer, not to each device.
+Hilal well it would fix it if everyone would switch over but the ISPs of the world really dont have any reason they should switch over. Its just expense for them, and it still works on ipv4 because of the patches we have with how routers work. Almost everything is still ipv4 but it hasnt broken because they can use routers to fake more addresses.
A=10, B=11, C=12, D=13, E=14, F=15, 10=16...19=25, 1A=26, 1B=27. IPv6 uses hexidecimal numbers, which is a base 16 number system, as opposed to our normal decimal system (base 10). It's worth mentioning that their both ultimately just a bunch of ones and zeros, and the only real difference is in the way that we have chosen to represent them for human reading.
+lisa lavergne +Brandon Phillips 2^128 - 2^64 = 340282366920938463444927863358058659840 IP6 64bit you see why yet? its MUCH bigger space of bytes to have.
Delete from the internet everything he has created a child under 13 years and you will get alot of terabytes (or zetabytes?) free space on whole internet
IP version 1= 4bit IP version 2= 8bit IP version 3= 16bit IP version 4= 32 bit IP version 5= 64bit IP version 6=128bit IP 2 was used experimentaly and the old version 2 IP packets were assigned to version 5 in order to avoid confusion with the normal IP adresses in version 4. Now the IP in version 5 that were handed out should not be around anymore, but to be sure to avoid confusion, we skipped 5.
Now Brady, I want you to make a video with James Clewett about if there is a possibilty of 86-bit computers being right around the corner. I know that there have been 64 bit computers since 2003 (I research), and by 3:00 A.M.. January 19, 2038, we are mostly going to have 64 bit computers. The challenge for you Brady is to find the maximum counting value of an 86 bit computer using math (which would be hard to decimalize since 64 bit is about 18.4 quintillion). I want you to record it, and post it on RUclips on one of your math channels. I really enjoy Numberphile and hope that you could do this without the help of anyone. That would be delightful.
the highest value for any x-bit number should also ALWAYS be an odd number, since the least significant bit (2^0) = 1 will be "on". this bit always determines whether a number is even or odd in any x-bit number
That makes a lot of sense. Also when considering how arrays are indexed. You can have an array with 256 positions and the highest index is 255, but there are still 256 items there. I should have known what he was talking about, I need to watch the video again to try and figure out why I thought there was some ambiguity in his statement and what confused me.
One slight correction: A switch doesn't look at the IP adress because it's usually a layer 2 device. It rather holds a database of MAC adresses to the respective port which would be a 48-bit number. There are "Layer-3-Switches" but they would be called a router by every IT person I know. Oh and please don't say "n-Bit computers can count up to 2^n", they can adress 2^n would be much more accurate because my C64 had no issues other than time and energy to count to many thousands with a for-loop.
3:55 Wrong. Any device that resides behind a router (that is if people want to have their oven, fridge, etc, hooked up to the internet) is communicating via NAT. They don't need their own external IP for comms. They only require an internal private IP and MAC address. But the need for IPv6 is apparent.
At my company, we're pushing back on IPv6 hard, as many of our mainframe apps would have to be nearly completely rewritten. The hard part isn't the routers or operating systems...it's the apps that have to communicate from behind the routers.
Who said anything would collapse? The point is, it's not a solution. It's a stopgap that will fail. It's already extremely taxing to keep all of this working. We want to move to a system that just works without having to make all these workarounds. It adds a high level of complexity. I also feel the need to point out that I'm talking NAT here, not just the basic subnet. A regular subnet doesn't help at all. You still need an WAN IP address for each computer.
Depending on one's situation, a router may or may not be a significant bottleneck. However, a router serves at least two other important functions: if properly configured, it filters out the majority of exploits such as port scans and malformed TCP packets so your devices never see them, and it creates an internal network where your devices have free access to one another and the internet, but your data is secure. Even if everyone switched to IPv6 tomorrow, I would not be without a router.
64bit is the register-size in the CPU which affects the size of operations you can do in a single tick of the CPUs internal clock. E.g. you can add two integers with the size of 64bit. This doesn't influence the representation of data. It's like you gave the postal service better calculators and expect that the postal code could address more cities. You can do this if you add more digits to the postal code (e.g. 4 digits=>10000, 5 digits=>100000), and this is basically what IPv6 does.
I did not know there was an actual place/company/thing that handed out IP addresses. Come to think of it, the thought of where they come from never occurred to me.
I imagine that they skipped IPv5 because in the past they had been incrementing by 8 (8, 16, 32) but then they decided to skip 64 bit IP addresses and went straight to 128 bit addresses.
Not all 4,294,967,296 IPv4 addresses are usable IPs (ex: 127.0.0.1). The reason why we haven't all switched yet is the same as why you can have 20 devices on your home network while only being issued 1 IP from your ISP: NAT, specifically PAT. Utilizing Private IP ranges and PAT and a well designed network architecture can allow a company to use only 1 to serveral IPaddr for their entire business. I'm pretty sure most home users are using some form of PAT/NAT without even possibly realizing it.
32-bit operating systems can handle 128-bit addresses, just like 16-bit operating systems and processors could do computations with numbers greater than 2^16, just not as efficiently. The 32- or 64-bitness refers to the widths of registers and buses, and larger numbers can be handled by combining more than one register and by sending them over the bus in parts.
Nope. It's just a sort of workaround. Sure, in theory, you'd get around 65,000 times as many addresses, but subnets also serve the function of isolating computers from other computers. You don't want a bunch of random computers on the same subnet.
Many websites can exist on the same IP. When you access a website, usually you connect to the server by the IP address and the server knows which website to give you by the host field.
1) A lot of number blocks are reserved, so the number would be much lower. 2) It's IANA that has run out of numbers. The problem is that once the regional authorities also run out of numbers, there are no more IPs to allocate to businesses. So if you wanted to start a new ISP... Well, tough luck.
Each router in these private networks remembers where to return information in it's address table after it receives a response. We would have run out of IP addresses a long time ago if every device had to have it's own IP address. There is a benefit to being able to give every device it's own address since there is some overhead in networking in this way, so IPv6 is going to be useful, but it is by no means a necessity. Which is why no one is in a big hurry to convert.
That's because they are mathematicians, and not computer scientists.. I as a computer science student in final year noticed the same mistakes and got a bit disappointed, but what can I say.. I think there should be a channel where computer scientists explain the most important things in their field.. This kind of video would belong there.. Like Numberphile is for math and Scishow for biology and other things.. We need an IT channel like those.
I agree, it's especially misleading when he gives the example of the computer, tablet and so on in your home. These things (almost) never have a unique external IP. They have local IPs.
An 8-bit computer can count to any number it has the total memory to store. You the first two bytes and remember the carry for the next two bytes. If you allocate 1024 bytes per number you get 8192 bits, which is a very big number. It's cumbersome, but doable. This sort of scheme would be reserved for special cases.
IPv5 wasn't 64bit. v5 had just some changes mostly in the routing handling and some other allocations of space inside the package but it was never used. IPv6 also upgrades in many more things than just addressing. It changes things from packet structure to packet handling in routers, it's just that the really pressing issue is the addressing.
Basically, it's because IPv6 isn't backwards compatible with IPv4. So pretty much the only way to be able to go from IPv4 to IPv6 without hassle or struggle would be to change to IPv6 software everywhere there's a computer at the same time. That and as he said, it's not economically appealing for companies to do that when they can just stick with IPv4 for as long as possible.
An ipv6 address is just an 128 bit number instead of a 32 bit number in the case of an ipv4 address. You can write this binary number in the hexadecimal number system which, as you noticed, uses letters too (digits 0-9 and "letters" A,B,C..,F) You can easily write any ipv4 address 'with letters' using a hexadecimal notation :) Ex. 231.18.45.20 -> 256^3*231 + 256^2*18 + 256*45 + 20 = 3876728084 (in decimal, E7122D14 in hex and in IPv6 style notation -> e712:2d14
Team Fortress has been being constantly updated for the last few years, constantly releasing new weapons, maps and accessories to keep it's fans happy. The last time we got a Half-Life game has been in 2007 and it ended with a cliff-hanger.
No, what he is saying is that the size of the IP address is not sufficiently big to provide a unique address to each machine connected to the Internet. But don't worry, because this problem has been identified a long time ago (1994 approx.) and we have been using tricks that avoid the problem. The real solution is the IPv6 address, which already has been implemented, but we are still migrating from IPv4 to IPv6.
In the UK, the body that was meant to be making this happen has folded (last week I think) when they realized that they were getting no support from the government and thus what they were trying to achieve was pointless. At the moment there is no UK government websites that use IPv6.
The biggest problem are actually the adoption of phones connected to the internet. The 3G phone network for example hands out none LAN IP addresses to each phone. This is because when you change between each phone tower you need to maintain the same IP address to make data transfer without interruption.
A lot of the larger service providers (with many of the large backbones) are already using IPv6. IPv6 is very difficult to adopt which is why it is not happening quickly at a micro level. What you are not taking into account is the private subnet ranges. For example the 1918RFC ranges: 10.0.0.0, 172.16.0.0 - 172.31.0.0 and 192.168.0.0 - 192.168.255.0 Source NAT enables you to use 1 IP to address many machines behind it, so it is not really the 1 to 1 mapping you are describing.
Some of the video is misleading. Early computers were 8-bit processors with 16-bit address lines. Usually, they had something like 4 kilobytes of memory. While any given register could only hold a value from 0 to 255, they were no more limited to that value than we are to the number 9. I can count higher than 9. I just use more than one digit to do so. Pac-man only used one byte for the level because space was at a premium and they thought no one would reach that level.
The only issue i see in this video (although it's correct) is the (without advanced configuration) 254 devices you can have in your home, do not use more then one IP from the 4,294,967,296 IP's. because your router hands out IP's from the 192.168.x.x (usually 1.x or 0.x) because thats your local network, not your wide area network. They failed to mention that in this video, so i feel people are misinformed about that. Who knows, maybe i'll make a video response and hopefully get them to reply.
He forgot to mention, that If it were'nt for Network Adress Translation (NAT) routing, we would have ran out of IP adresses somewhere from 10 to 15 years ago. Although NAT had been invented for security purpouses, it has been extensively used for decades to connect private networks to the "outside" using only one IP adress. And of course - web servers usually host a multiple number of websites, each with it's unique domain name, under a single IPv4 adress. Personally, I still prefer to avoid using IPv6 unless really neccesary. Mostly because IPv6 is considered as poorly implemented in Windows (there are some known security and stability issues), it is not backwards compattible (although it could have been designed as such), and overall I think, that the engineers who proposed and designed IPv6 did a lousy job.
Routers use NAT to reduce the number of addresses used. The addresses on your local network can be whatever you want them to be, and your router is given one (or more) address(s) that will encompass your whole network. This video is slightly misleading on some topics. Source: I am a Cisco Certified Network Associate.
The maximum value for a *signed* 32 bit integer is 2,147,483,647, but there are another 2,147,483,646 numbers that are less then zero. This makes the total number of numbers 4,294,967,296.
...But local addresses on a network are only local addresses meaning that anyone could have say 192.168.1.1 on multiple networks.. what he needed to say is that the internet IP provided by their ISP's (Internet Service Provider) are running out, so I don't think that is necessarily correct for a university scale internet structure as they might use 2 or 3 (maybe more) internet connections or different IP's given by their ISP or even enterprise grade internet structures which allows them to have more connected users to a single modem than the typical home user would.
...Well, no, because hunting a John with the last name Conner would have sent him on a goose-chase while John Connor keeps playing arcade games and being generally gangsta.
your IPv5 is called Insternet Stream Protocol. Never was released on consumer level. Wasn't called IPv5, it's just that the version position in the header was flagged as '5' to distinguish it from the other protocols.
Be pretty funny if the video just ended at 2:14
+JessLe Berry funnier at 5:36
+Nemo's Channel Indeed, LOL.
+Nemo's Channel Even better at 8:19
+Jakob7539511 I see what u did thar.. Which means you made me look >:|
:3
"I like the us defense department, they are good people" close one, but you covered your butt there.
NO
the biggest 32-bit integer is 2^32-1
the *number of values* is 2^32
the first value is 0
but even that is not the whole picture, because for instance all IPs that start with 192.168 are reserved for private use and cannot be given out nor should be used outside a LAN
and there are other reserved numbers
another aspect of the problem is that addresses are given out in blocks which are hardly ever completely used
the same problem will possibly confront the world's telephone systems someday
but no worries, the IT industry is, as was said, in the process of switching to 128-bit IPs
128 bits is enough to give every electron in our galaxy its own address
assuming we dont waste too many addresses
OMG. Is this what physicists feel like when they read the comments under physics vidoes ?
Back in the days before the web I would often remember the straight IP address instead of a domain name. Only for places I used regularly though. These days I don't even know what my own IP address is.
Allan Richardson No I am not paying extra. My ISP was the first one ever in the UK and they were pretty innovative developing new technologies which most other later ISP's adopted. They were true techies. They offered an excellent service without relying on irrelevant bells and whistles, or questionable business practices such as AOL.
Unfortunately they were eventually bought by a larger ISP, and once the suits took over, the quality of the service started to drop. Tech support used to be by Brits who really knew what they were talking about. Now it's been outsourced to India where you get some random person with a hard to understand accent reading through a script.
Technically the 32 bit scheme is infinitely extensible by using NAT. Also, the video acts like the devices in your house take up extra adresses, but that is not true. They will be off of your single home IP, and internally addressed with at 10.x or 192.x address
Not infinitely, NAT extends IP addresses by combining IP address and port number to uniquely identify a computer. Port numbers are limited to 65,535 (16 bits). So NAT can extend 32bit addressing to at most 48bit, usually much less because normal work on a computer (such as watching RUclips) needs a dozen or two of port numbers.
Why infinitely extensible? I guess it's about 72,000,000,000,000,000 total addresses with NAT
Was the problem that Internet Servers needed unique IP addressing? Most Net users are normally clients and dont need the full address exclusivity, just a unique, dynamic or translated return path. Was IPv5 going to be a 64 bit system, but the technology jump to 128 came in faster than IPv5 could be commercially developed
@@stensoft more NAT
Listen carefully at 5:34 and you can hear helicopter blades.
IoEstasCedonta The US defence department heard he doesn't like them
Nitay A. He did conclude. "No, they are nice people... Oh dear." Hopefully not "famous last words":)
nimim. Marko Mikkilä I do believe that the Defense Department handed over full control of the Internet to five regional consortia of universities in the U.S. quite some time ago, now. It is they that own the backbone, in any case, and I don't believe that the Defense Department is paying for any of it at this point. So Professor Clewett can relax!
As for his wondering about domain names, they were instituted by the same team that developed IP and have always been part of the Internet.
Those are computer fans just saying. I mean it's not like theres an open computer lying right next to them.
I was waiting for him to mention Network Address Translation (NAT), but he didn't.
Steven Titus For website hosting on a dedicated address without depending on a centralized proxy (which NAT based upon), you need to buy the right to a IPv6 address from the IP address managing group (name was in the video, but I don't remember off the top of my head), or you need to find a willing seller of a IPv4 address holder. Market Pricing via supply and demand is a temporary and decentralized solution to this as you, but if the expansion of the Internet of Things keeps up its exponential pace, then the demand will overwhelm the supply of sellers to the point where prices of IPv4 addresses will skyrocket, thus forcing individuals and companies to either accept the pyramid like structure and centralization that NAT forms (which we somewhat have now), and/or buy the rights to never used IPv6 addresses.
Big ISPs like comcast already use NAT for their customers. All of the comcast customers in my area use the same public IPv4 address from the regional hub, so then from the centralized hub, every customer is given a unique IPv4 address from the viewpoint of the central hub via NAT as an intranet. All router-based LAN networks work the same way - all devices have their own IPv4 (and usually IPv6 address as well) from the view point of the router as a intranet, and connect to the rest of the internet via NAT on your router and your modem.
Kyle Herbig Yes, I know how NAT works. I was just making an observation that he didn't mention it in the video. At 2:16 he could have mentioned NAT. Its just a little bit misleading, that's all i'm saying.
Steven Titus
I agree - I thought you knew, so I was writing for the casual comment reader. ;)
+Steven Titus Yea, I too thought it was a bit strange to leave that out. Since the statement that all your devices need a unique address just isn't true.. I even heard from people in India that they actually got internet via some sort of LAN network with their ISP. Which, although not a great solution, is a solution to the problem of running out of addresses. (Btw I understand that in the end we just need ipv6 )
It's not a bad solution either, you can stuff 16 million machines in the class A private range
Will Pegasus Galaxy ever run out of stargate adresses?
+Кэтишкэт Старье 7 number addresses + a point of origin. And if i counted correctly there are about 40 symbols on each stargate - 1 for the point of origin. Without all the working its about 15.4 million 15,380,937. Pegasus ( at least in real life ) is a dwarf galaxy and it only has a few billion stars, so lets be generous and say 10 billion. Using the parts of the drake equation we need it comes out at about 3.2 million planets that can support human like life. So no there are plenty.
Andrew Joy Nice explanation, but I hope you did get the joke.
+Andrew Joy there is a 9th spot to denote a moving ship
How does that work ? no point of origin and no fixed address as the ship moves!
Andrew Joy 9th Chevron is dependent on distance between Destiny and Earth (SG:U).
Cheers to anyone watching on 64 bits (in 2016)=]
Cheers
GodlikeBlock which is nearly everyone -_-
so does 64x mean 64 bit and 32x mean 32 bit then?
OrangeCreeper217 Yesh i think
OrangeCreeper217 no duh?
Your refrigerator is made by "SMEG", I don't think I could own that fridge without laughing every time I see it.
What's SMEG?
@@keval2137 Please see Red Dwarf, the other definition may be on an on-line slang dicktionary. Wallace and Grommit's one was called Smug.
03:47
Simple solution:
IPv6
Isaac Karjala To be honest, about a decade ago I would have agreed with you. There was a point in time where I really wanted to have every computer on my home network with a routable address for protocols like you mentioned.
But the way that so many Internet-connected home devices these days work, you go through some third-party portal site that interacts with the device using a proprietary protocol. My Nest thermostats (yeah, I have two in my house for different zones), for instance, are stuck behind NAT but the Nest website straightens all that out for me.
Western Digital has a sort of similar service for their network attached storage devices. It's like the future is to hide protocol details behind some cloud service. I hate the phrase "Internet 2.0", but that really seems appropriate ;)
Simple solution:
_watch the whole video_
Simple answer:
_I did_
Connor Wright
Simple reply:
_why suggest it then?_
Simple, but unrelated comeback:
_Because you're adopted_ hahahah
I like IPv4 as I can remember it easier but we need IPv6 just wish there was an easier way to remember HEX
What? I was typing a comment and RUclips just went to the next video?
I was saying there are fewer than 4,294,967,296 IP addresses in use on the Internet. All the RFC 1918 addresses, as well as numbers beginning with 0 ("Discovery"), 127 ("Loopback") and 224-255 (Multicast and "Reserved") make up a big chunk of addresses you'll never see, along with a bunch of others. My calculation that is 592,708,864 addresses (13.8% of the number given) not usable. This doesn't even include address ranges like 1.1.1.x and 1.2.3.x that are informally reserved, plus some companies and the DOD have large blocks that may not actually be in use.
its actually more complicated than that most homes get one ipv4 address and each device gets a non routable address behind that address much like how apartment numbers work
This is supposed to change with IPv6 though, isn't it? With the incredible amount of available addresses, basically every device in a household should be able to get a public address themselves.
more like a billion addresses lol
+SirLugash That would be a ton less secure, I genuinely hope that's not the case lol
Wade Stutzman
Well that's what I've read. We'll see.
no less or more secure. saying its non routable is not 100% true. My understanding is the outer header for the packet contains the ipv4 address for your house, then your local address is at the next stage of the header. you can start pinging for the local adresses, guessing at local routing names (usually people will have 192.168.0.1 or some other std format), and find the full address for a computer behind a local network. Also most attacks originate from the computer itself in the form of malware, in that case you already know the address.
also I don't know the ratio, but most budget internet connections do not assign a unique ip address to each home, instead sharing them by assigning a dynamic IP address each time you reset your connection (and sometimes while its still running). usually doesn't cause an issue, but you can't run a home server accessible offsite, or be the server for an online game, or home host a website with a dynamic ip (easily).
switching to ipv6 would make things alot easier on many levels, and in reality the added layer of obscurity in ipv4 is only an annoyance to an attacker, not something that significantly reduces the occurence of attacks.
one thing u have to remember is that in any x-bit number its always going to be (2^x) - 1, not just 2^x to get the highest possible value
(2^(x-1))-1 if it's signed, even. Fortunately, your IP address has no need of being signed seeing as we don't expect to address any negative devices.
Man, I never knew Heston Blumenthal knew this stuff.
I remember in the 1990s we thought that these would never be used up.
4,294,967,296 may look like some crazy random number, but It converts to 100000000000000000000000000000000 in binary.
It is actually the 4,294,967,286th in a series that starts at zero , but by convention we start counting at 1, not zero. So it rolls over to the next place even though the next register doesn't exist in the computer. It gets to 4,294,967,296 because we talk about it as if we started counting at 1..
Another example is when a dollar equals 100 cents - where this takes three places.
But you can count to 100 if you start at zero as the first, and 99 as the hundredth. That only requires two places.
That's the difference between computers, where IP tables are maintained, and everyday reckoning. And yes, iIt can be confusing.
4,294,967,295 would be 11111111111111111111111111111111
It's not confusing if you ask proper question.
How many 1-digit numbers can you write?
10 - first 2-digit number.
How many 2-digit numbers can you write?
100 - first 3-digit number.
...
How many 32-digit numbers can you write?
1 0000 0000 0000 0000 0000 0000 0000 0000 - first 33-digit number.
And it's true for any system as long as 0 mean zero and 1 mean one :)
Carpey computers count from 0, and we count from 1, e.g. the first number would be 0000000 in 8 bit, if you count from 0 it would be 11111111111111111111111111111111
which converts to a single bit in -24h clock notation-
There you go ladies and gentlemen, the United States Defense Department, the creators the the internet.
That's not what he said... They govern the IP tables...
The origins of the internet is TCP/IP (the protocol that every computer is using to talk to each other) and that was created in CERN in Switzerland.
OverSoft Well, the technology of the internet was created by the United States Defense Department(or at the very least was funded by it), but the WWW was created at CERN.
To be complete the WWW was created on a NeXT computer, btw (which is now Apple's Mac OSX).
TheGreatRakatan They're not the creators of the internet, they just determine public ip adresses. The first people that connected computers to communicate with each other and they've called it net were from some California university. I'm sorry, if I'm wrong, but this is what i remember from the CNA (Cisco Networking Academy) interpretation.
TheGreatRakatan CERN created the internet to link up their computing grid I believe.
You are correct if you start to count at one. His main point was that an 8-bit number has 256 different values, i.e., 0 to 255. If you encode binary numbers in two's complement, then an 8-bit number would range from -128 to 127 (so can only count to 127 in this case), but still have 256 different values.
And now we use 64-bit. We still advance. And what about internal IP's?
It's not that easy. If You have only internal IP's then You don't have Your own external IP.
+Byamarro2 Well internal (or private) IP's are a HUGE contributor to conserving IP addresses, as essentially every household need only 1 external IP address (companies might have a few, as they might have a few different connections outbound but still, likely less than 10 for hundreds if not thousands of machines)
I know this video is 4 years old but even then most people viewing these videos, at least on PC, were using 64 bit computers.
How many individual IP's could we have if we scraped numerical or hex and replaced it with alphabetical. I.e A=1 - Z-26?
I learn so much through Numberphile. I love it.
Brady, you asked about content for the new computing channel: James has covered key elements of domain and ip fundamentals really well there. Maybe some more of this type of content?
At 4:51, there is a typo in the caption/subtitle: it should be DHCP, not DHTP
4:50 caption error. Should be "DHCP," not "DHTP"
Why they decide to use 128 bit instead of 64?
64 bits it already 18 quintillions of addresses.
But to fill 128bit you need to convert all mass of the earth into devices, and each one should weight 5*10^(-11) gram!
That is ridiculous!
Right... 64bit seems enough.
They probably want to avoid the same problem in the future. Humans are notoriously bad at predicting the future, people said the car would never replace horses, no one would ever need a computer in their house, people would never need more than a megabyte of memory etc. Back in the 80s and 90s they probably figured that there would never be billions of computers using IP addresses or that by that time computers would have advanced past the severe processing limitations they had then. Given the proliferation of computer technology who knows how many devices will require IP addresses in 50 years or 100 years etc? Better to just make it 128-bit and never worry about it again, its not like its too demanding in processing power. There are also more technical and security reasons as well
whos to say how the future will work like thereine said who knows what will use them perhaps nano machine will use them up or small drones best go with the big number
But that's only planet earth! So small! What about the entire galaxy? Or bigger? Sure we may not get there, but if we do, it'll be well worth the investment. Also, imagine fixing 18 quintillion devices to support IPv24658. Doesn't sound very easy, does it?
What most forget is that a single device might use multiple addresses at the same time. There are multicast or unicast addresses which devices can "use" to actually hook into a stream. That's already happening with IPv4, though there are just a few multicast addresses out there ^^
There are a lot of special addresses which are locked for special uses. For example the IPv4 reserves a whole class A subnet just for the local loopback address. This address can only be used locally on each device and always refers to the device itself. The whole 127.X.X.X space is wasted for that ^^. That's 16 million addresses or 1/256.of all addresses in the IPv4 address space.
It would also simplify the assignment of IP addresses. You could actually define certain addresses to be bound to the location on earth. So each square meter might have a million addresses available. A seperate address space might be used for mobile devices (as they very likely change position).
The GUID also uses 128 bits. It's "said" that the possibility that the same GUID is picked twice is very very unlikely. Same for MD5 hashes, they are also 128bits long. Since modern computer hardware is already able to natively handle 128bit values it's no suprise they picked 128bits for IPv6.
Network Address Translation
The local network they run on shares the same IP to the router, and then the router redirects the traffic to the appropriate internal address. If you want a more accurate description we have wikipedia for that.
Numberphile's IP address is 109.123.72.70
I seriously doubt I could ever remember that. Praise the people who developed Domain Names!
IPv6 anyone....?
6:00
Lukasz Baldyga I know that was my thought through this whole thing
lol the US defense department controls the internet
this should be posted on computerphile
If people are still wondering. Domain Name is basically the 'mask' of its IP. Then a DNS is used to translate the Domain Name to its IP Address, in other words 'see what's behind the mask'. Think of it as a Phonebook. Doman Name is the Name. IP Address is the Number. DNS (Domain Name Server) is the Phonebook itself. Very brief and vague explanation but I tried to keep it simple.
when they say, "computer can only count up to 4 billion" they mean when it specifically comes to this address thing. the thing is, those old computers can only handle numbers up to 4 billion at once, which it needs to do for addresses. but when it comes to stuff at the user level, such as a calculator or storing the number of youtube views, it can take the number in chunks, and have no problem handling larger numbers.
Apparently RUclips was forced to update their view count system to 64 bit since Gangnam style got over 2,147,483,647 views, which is the highest possible number in a signed 32 bit system. Maybe they should have went with unsigned integers to hold up to 4,294,967,295 views. It's not like there's negative views.
Well fast forward 3 years and Despacito has more than 4,294,967,295 views. So what is their new number?
9 sexdecillion i think
Better yet, you can do subneteting much easier than with friggin IPV4.
+ComandanteJ that's a solution we're using rn :)
Lol, not really.
Source: CCNP
ComandanteJ The only difference is there is no classful boundary or broadcast addresses and hex characters versus decimal. I guess the difficulty of the process is debatable. I found it to be about the same since your main concern in both ipv4 and ipv6 is how many bits represent the subnet bits and how many represent host bits. In IPv6 you are converting to binary from hex instead of decimal. In my opinion the difficulty of the process is the same.
This guy's talking as if every device gets a different public IP. Routers give local IPs and then the modem is the only device getting a public IP.Please slap that fool.
Doesn't matter. We're still running out because of the number of servers.
I never said we weren't.
Including public/private address spaces and mechanisms like NAT in this video would just add confusion on the topic. It also doesn't help since the problem described still applies.
He could've mentioned that in a regular home, it really doesn't matter how many devices you use, because your router is probably set up as a NAT, meaning it counts as 1 IP globally.
NAT is still used by almost everyone. The good thing is that it's 99% automatic, so no one notices it. Any time you connect your computer to a wifi hotspot, or a router at home or work, that router gives it an address in the 192.168.x.x block. The router then uses NAT to let you talk to the outside world.
This means the ISPs only have to give one unique IP address to each customer, not to each device.
didnt ipv6 solved this problem?
yea realized that later,,
+Hilal Just what I was thinking.
+Hilal well it would fix it if everyone would switch over but the ISPs of the world really dont have any reason they should switch over. Its just expense for them, and it still works on ipv4 because of the patches we have with how routers work. Almost everything is still ipv4 but it hasnt broken because they can use routers to fake more addresses.
6:00
That makes sense I was watching this video and wondering... isnt my ip (Home IP Address):(Device #)
3:48 SMEG
"Bitesize" lol
+Andrew Shirley I caught that too!
Such a nice person to listen.
A=10, B=11, C=12, D=13, E=14, F=15, 10=16...19=25, 1A=26, 1B=27.
IPv6 uses hexidecimal numbers, which is a base 16 number system, as opposed to our normal decimal system (base 10).
It's worth mentioning that their both ultimately just a bunch of ones and zeros, and the only real difference is in the way that we have chosen to represent them for human reading.
I have a 64 bit ._.
Brandon Phillips so does the majority of people now but is you router on IPV6?
+lisa lavergne +Brandon Phillips
2^128 - 2^64 = 340282366920938463444927863358058659840
IP6 64bit
you see why yet?
its MUCH bigger space of bytes to have.
why can not just allocate all the ipv4 adresses as a subdomain of ipv6 :P
My N64 has 64 bits
The video is wrong in so many ways that it's very painful to watch.
do enlighten us
How so?
Delete from the internet everything he has created a child under 13 years and you will get alot of terabytes (or zetabytes?) free space on whole internet
It's not space for things we're having not enough of, it's not enough of garages where we can store these things!
32-bit signed integers range from -2,147,483,647 to 2,147,483,647 while signed integers range from 0 to 4,294,967,296.
IP version 1= 4bit
IP version 2= 8bit
IP version 3= 16bit
IP version 4= 32 bit
IP version 5= 64bit
IP version 6=128bit
IP 2 was used experimentaly and the old version 2 IP packets were assigned to version 5 in order to avoid confusion with the normal IP adresses in version 4. Now the IP in version 5 that were handed out should not be around anymore, but to be sure to avoid confusion, we skipped 5.
Now Brady, I want you to make a video with James Clewett about if there is a possibilty of 86-bit computers being right around the corner. I know that there have been 64 bit computers since 2003 (I research), and by 3:00 A.M.. January 19, 2038, we are mostly going to have 64 bit computers. The challenge for you Brady is to find the maximum counting value of an 86 bit computer using math (which would be hard to decimalize since 64 bit is about 18.4 quintillion). I want you to record it, and post it on RUclips on one of your math channels. I really enjoy Numberphile and hope that you could do this without the help of anyone. That would be delightful.
the highest value for any x-bit number should also ALWAYS be an odd number, since the least significant bit (2^0) = 1 will be "on". this bit always determines whether a number is even or odd in any x-bit number
That makes a lot of sense. Also when considering how arrays are indexed. You can have an array with 256 positions and the highest index is 255, but there are still 256 items there. I should have known what he was talking about, I need to watch the video again to try and figure out why I thought there was some ambiguity in his statement and what confused me.
I just want to view life with the same enthusiasm he discusses this subject.
One slight correction: A switch doesn't look at the IP adress because it's usually a layer 2 device. It rather holds a database of MAC adresses to the respective port which would be a 48-bit number. There are "Layer-3-Switches" but they would be called a router by every IT person I know. Oh and please don't say "n-Bit computers can count up to 2^n", they can adress 2^n would be much more accurate because my C64 had no issues other than time and energy to count to many thousands with a for-loop.
Just because the majority thinks it, does not mean that it is the right way of doing things
3:55 Wrong. Any device that resides behind a router (that is if people want to have their oven, fridge, etc, hooked up to the internet) is communicating via NAT. They don't need their own external IP for comms. They only require an internal private IP and MAC address.
But the need for IPv6 is apparent.
And I'm nerd enough to have checked to be sure that was an accurate translation. Well played, Ninja.
At my company, we're pushing back on IPv6 hard, as many of our mainframe apps would have to be nearly completely rewritten. The hard part isn't the routers or operating systems...it's the apps that have to communicate from behind the routers.
Who said anything would collapse? The point is, it's not a solution. It's a stopgap that will fail. It's already extremely taxing to keep all of this working. We want to move to a system that just works without having to make all these workarounds. It adds a high level of complexity.
I also feel the need to point out that I'm talking NAT here, not just the basic subnet. A regular subnet doesn't help at all. You still need an WAN IP address for each computer.
Depending on one's situation, a router may or may not be a significant bottleneck.
However, a router serves at least two other important functions: if properly configured, it filters out the majority of exploits such as port scans and malformed TCP packets so your devices never see them, and it creates an internal network where your devices have free access to one another and the internet, but your data is secure.
Even if everyone switched to IPv6 tomorrow, I would not be without a router.
3:47 that fridge reminds me of a Marshmallow XD
64bit is the register-size in the CPU which affects the size of operations you can do in a single tick of the CPUs internal clock. E.g. you can add two integers with the size of 64bit. This doesn't influence the representation of data.
It's like you gave the postal service better calculators and expect that the postal code could address more cities. You can do this if you add more digits to the postal code (e.g. 4 digits=>10000, 5 digits=>100000), and this is basically what IPv6 does.
I enjoyed decrypting these obscure messages. I now feel oblivious.
I did not know there was an actual place/company/thing that handed out IP addresses. Come to think of it, the thought of where they come from never occurred to me.
And now I'm shaking my head over the failure to specify the difference between machines and modems. Feel free to correct me if I'm wrong, though.
The biggest value you can store in a 32-bit number is actually (2^32)-1. 2^32 is the number of values you can store in it.
I imagine that they skipped IPv5 because in the past they had been incrementing by 8 (8, 16, 32) but then they decided to skip 64 bit IP addresses and went straight to 128 bit addresses.
They didn't skip it, it does exist, but it never became an official protocol
Not all 4,294,967,296 IPv4 addresses are usable IPs (ex: 127.0.0.1). The reason why we haven't all switched yet is the same as why you can have 20 devices on your home network while only being issued 1 IP from your ISP: NAT, specifically PAT.
Utilizing Private IP ranges and PAT and a well designed network architecture can allow a company to use only 1 to serveral IPaddr for their entire business. I'm pretty sure most home users are using some form of PAT/NAT without even possibly realizing it.
32-bit operating systems can handle 128-bit addresses, just like 16-bit operating systems and processors could do computations with numbers greater than 2^16, just not as efficiently. The 32- or 64-bitness refers to the widths of registers and buses, and larger numbers can be handled by combining more than one register and by sending them over the bus in parts.
Nope. It's just a sort of workaround. Sure, in theory, you'd get around 65,000 times as many addresses, but subnets also serve the function of isolating computers from other computers. You don't want a bunch of random computers on the same subnet.
Many websites can exist on the same IP. When you access a website, usually you connect to the server by the IP address and the server knows which website to give you by the host field.
I'll just put this an easy way: yes, it's to make it easier. Memory allocation is best if each upgrade is divisible by 2 and 8.
1) A lot of number blocks are reserved, so the number would be much lower. 2) It's IANA that has run out of numbers. The problem is that once the regional authorities also run out of numbers, there are no more IPs to allocate to businesses. So if you wanted to start a new ISP... Well, tough luck.
Each router in these private networks remembers where to return information in it's address table after it receives a response. We would have run out of IP addresses a long time ago if every device had to have it's own IP address. There is a benefit to being able to give every device it's own address since there is some overhead in networking in this way, so IPv6 is going to be useful, but it is by no means a necessity. Which is why no one is in a big hurry to convert.
That's because they are mathematicians, and not computer scientists.. I as a computer science student in final year noticed the same mistakes and got a bit disappointed, but what can I say.. I think there should be a channel where computer scientists explain the most important things in their field.. This kind of video would belong there.. Like Numberphile is for math and Scishow for biology and other things.. We need an IT channel like those.
I agree, it's especially misleading when he gives the example of the computer, tablet and so on in your home. These things (almost) never have a unique external IP. They have local IPs.
An 8-bit computer can count to any number it has the total memory to store. You the first two bytes and remember the carry for the next two bytes. If you allocate 1024 bytes per number you get 8192 bits, which is a very big number. It's cumbersome, but doable. This sort of scheme would be reserved for special cases.
Thought the same thing. But in the end it only prolongs the actualy problem instead of solving it.
IPv5 wasn't 64bit. v5 had just some changes mostly in the routing handling and some other allocations of space inside the package but it was never used. IPv6 also upgrades in many more things than just addressing. It changes things from packet structure to packet handling in routers, it's just that the really pressing issue is the addressing.
Basically, it's because IPv6 isn't backwards compatible with IPv4. So pretty much the only way to be able to go from IPv4 to IPv6 without hassle or struggle would be to change to IPv6 software everywhere there's a computer at the same time. That and as he said, it's not economically appealing for companies to do that when they can just stick with IPv4 for as long as possible.
An ipv6 address is just an 128 bit number instead of a 32 bit number in the case of an ipv4 address. You can write this binary number in the hexadecimal number system which, as you noticed, uses letters too (digits 0-9 and "letters" A,B,C..,F)
You can easily write any ipv4 address 'with letters' using a hexadecimal notation :)
Ex. 231.18.45.20 -> 256^3*231 + 256^2*18 + 256*45 + 20 = 3876728084 (in decimal, E7122D14 in hex and in IPv6 style notation -> e712:2d14
Team Fortress has been being constantly updated for the last few years, constantly releasing new weapons, maps and accessories to keep it's fans happy. The last time we got a Half-Life game has been in 2007 and it ended with a cliff-hanger.
I honestly actually feel embarrassed that I asked that question.
No, what he is saying is that the size of the IP address is not sufficiently big to provide a unique address to each machine connected to the Internet. But don't worry, because this problem has been identified a long time ago (1994 approx.) and we have been using tricks that avoid the problem. The real solution is the IPv6 address, which already has been implemented, but we are still migrating from IPv4 to IPv6.
In the UK, the body that was meant to be making this happen has folded (last week I think) when they realized that they were getting no support from the government and thus what they were trying to achieve was pointless. At the moment there is no UK government websites that use IPv6.
The biggest problem are actually the adoption of phones connected to the internet. The 3G phone network for example hands out none LAN IP addresses to each phone. This is because when you change between each phone tower you need to maintain the same IP address to make data transfer without interruption.
A lot of the larger service providers (with many of the large backbones) are already using IPv6. IPv6 is very difficult to adopt which is why it is not happening quickly at a micro level.
What you are not taking into account is the private subnet ranges. For example the 1918RFC ranges: 10.0.0.0, 172.16.0.0 - 172.31.0.0 and 192.168.0.0 - 192.168.255.0
Source NAT enables you to use 1 IP to address many machines behind it, so it is not really the 1 to 1 mapping you are describing.
Because its virtually irrelevant to the problem at hand. And, also, IPV6 is a way more elegant approach to hardware identity.
Some of the video is misleading. Early computers were 8-bit processors with 16-bit address lines. Usually, they had something like 4 kilobytes of memory. While any given register could only hold a value from 0 to 255, they were no more limited to that value than we are to the number 9. I can count higher than 9. I just use more than one digit to do so. Pac-man only used one byte for the level because space was at a premium and they thought no one would reach that level.
The only issue i see in this video (although it's correct) is the (without advanced configuration) 254 devices you can have in your home, do not use more then one IP from the 4,294,967,296 IP's. because your router hands out IP's from the 192.168.x.x (usually 1.x or 0.x) because thats your local network, not your wide area network.
They failed to mention that in this video, so i feel people are misinformed about that. Who knows, maybe i'll make a video response and hopefully get them to reply.
He forgot to mention, that If it were'nt for Network Adress Translation (NAT) routing, we would have ran out of IP adresses somewhere from 10 to 15 years ago. Although NAT had been invented for security purpouses, it has been extensively used for decades to connect private networks to the "outside" using only one IP adress. And of course - web servers usually host a multiple number of websites, each with it's unique domain name, under a single IPv4 adress.
Personally, I still prefer to avoid using IPv6 unless really neccesary. Mostly because IPv6 is considered as poorly implemented in Windows (there are some known security and stability issues), it is not backwards compattible (although it could have been designed as such), and overall I think, that the engineers who proposed and designed IPv6 did a lousy job.
Routers use NAT to reduce the number of addresses used. The addresses on your local network can be whatever you want them to be, and your router is given one (or more) address(s) that will encompass your whole network. This video is slightly misleading on some topics. Source: I am a Cisco Certified Network Associate.
Its 2015, my cellphone is 64-Bit now. The technological advance over the last 3 years !
like he said, lots of people aren't updating, cause not just anyone can, and a lot of people don't really know about this kind of stuff
The maximum value for a *signed* 32 bit integer is 2,147,483,647, but there are another 2,147,483,646 numbers that are less then zero. This makes the total number of numbers 4,294,967,296.
Where did you get a "SMEG" refrigerator? I'm in awe
...But local addresses on a network are only local addresses meaning that anyone could have say 192.168.1.1 on multiple networks.. what he needed to say is that the internet IP provided by their ISP's (Internet Service Provider) are running out, so I don't think that is necessarily correct for a university scale internet structure as they might use 2 or 3 (maybe more) internet connections or different IP's given by their ISP or even enterprise grade internet structures which allows them to have more connected users to a single modem than the typical home user would.
...Well, no, because hunting a John with the last name Conner would have sent him on a goose-chase while John Connor keeps playing arcade games and being generally gangsta.
your IPv5 is called Insternet Stream Protocol. Never was released on consumer level. Wasn't called IPv5, it's just that the version position in the header was flagged as '5' to distinguish it from the other protocols.