network and host bits in a ip

[ilyas001]

well guys  i try to understand up technologies here for my certification so i will ask questions even if they are totally wrong just to let you know how i think so you would be able to understand where i miss understood something and be more able to help me because i'm on a phase where i uderstand little so little i don't even know what to ask lol

so i don't get it we have the ipv6 it's big enought so you don't need your modem to use NAT  technique de get to the internet so i have a question . if there are so many internet protocols possible why is that there are still private and public ips even with ipv6 

and if ipv6 don't need nat technique i can go directlly to the interent because it have a network and host part right ? then why ipv4 have and network part in it if anyway it's gonna be translated by nat to go over the internet for what purpose we have that network part in ipv4

thanks for your time 

[Wombo]

2 hours ago, ilyas001 said:

well guys  i try to understand up technologies here for my certification so i will ask questions even if they are totally wrong just to let you know how i think so you would be able to understand where i miss understood something and be more able to help me because i'm on a phase where i understand little so little i don't even know what to ask lol

so i don't get it we have the ipv6 it's big enought so you don't need your modem to use NAT  technique de get to the internet so i have a question . if there are so many internet protocols possible why is that there are still private and public ips even with ipv6 

and if ipv6 don't need nat technique i can go directlly to the interent because it have a network and host part right ? then why ipv4 have and network part in it if anyway it's gonna be translated by nat to go over the internet for what purpose we have that network part in ipv4

thanks for your time 

There's aren't really "private" IPs in IPv6.

 

You have Link Local which are used for local communication and will be automatically assigned by the host even if there is no subnet assigned or DHCP available. This allows for easy plug-and-play when it comes to IPv6, as far as local networks are concerned anyway. These addresses are always used for communication over the local subnet and will always be in the FE80::/10 range.

 

Site Local addresses are what most people refer to as the private addresses of IPv6, but these are highly contested and are by no means necessary, or even required. These addresses have been added and removed from the standard many times, and the debate is still up on their validity within the standard. Essentially when coming from IPv4 to IPv6 the notion that we wouldn't use NAT was scary to some. This forced the adoption of a quasi private address space in IPv6 however unless you go out of your way to use these addresses and implement NAT on top of them they can basically be ignored. They serve no inherit purpose within the standard aside from fulfilling some people's inherent want/need for NAT that does not need to exist in the IPv6 world. These addresses will always be within the FD00::/8 range, which is part of the FC00::/7 subnet space currently allocate for site local, or "unique local" addresses. The only currently assigned block (for private use) from this space is FD00::/8 however.

 

Lastly your Global Unicast addresses are your real, routable, IPv6 address, with Link Local and Site Local being non-routable. In other words, Global Unicast addresses are the only address that are routable on the open internet and allow for public inter-network communication. The current allocated space for Global Unicast addresses is 2000::/3. Addresses within this space are allocated via regional IP authorities, just link public IPv4 addresses. Allocation of these addresses are restricted to minimum allocations of /48. Anything smaller must come from a local LEC, carrier, or other institution who will "lend" you some of their assigned address space, much like how your current ISP lends you one of their IP's so you can reach the internet.

 

It is important to note that while you can subnet IPv6 addresses below /64, ie. /123 for wan links, this is not recommended for larger network segments as lowering the subnet to blow the 64th bit will cause EUI64, and the variants thereof, to stop functioning, requiring manual assignment of IPv6 addresses or the use of DHCP for IP assignment.

 

The main reason we use NAT in IPv4 is for address preservation. There are far too many devices that need IP addresses and nowhere near enough IPv4 addresses to go around. This lead to the wide-spread adoption of NAT which up until that point, was more-or-less only used in specific situations requiring the obfuscation of IP's and a few other fringe usecases. In the IPv6 world we have more than enough addresses for everyone to have millions, so we just don't need to use NAT in the IPv6 world.

 

NAT in IPv6 does serve an important function in inter-compatibility with IPv4 however in the future when everything is IPv6, this will no longer be a factor.

 

As for the network and host bits of address, they are just used to denote which parts belong to the network (assigned by regional authorities and controlled) and the host bits, which are entirely user controlled. Combined with the subnet mask any computer knows instantly if the address it is trying to reach is part of its own local subnet (network bits matching) or if it is outside its local subnet and requires its default gateway to reach it. Without subnet masks, network bits and host bits, our computers wouldn't have a way to know where they need to send their network traffic.

[ilyas001]

4 hours ago, Wombo said:

There's aren't really "private" IPs in IPv6.

 

You have Link Local which are used for local communication and will be automatically assigned by the host even if there is no subnet assigned or DHCP available. This allows for easy plug-and-play when it comes to IPv6, as far as local networks are concerned anyway. These addresses are always used for communication over the local subnet and will always be in the FE80::/10 range.

 

Site Local addresses are what most people refer to as the private addresses of IPv6, but these are highly contested and are by no means necessary, or even required. These addresses have been added and removed from the standard many times, and the debate is still up on their validity within the standard. Essentially when coming from IPv4 to IPv6 the notion that we wouldn't use NAT was scary to some. This forced the adoption of a quasi private address space in IPv6 however unless you go out of your way to use these addresses and implement NAT on top of them they can basically be ignored. They serve no inherit purpose within the standard aside from fulfilling some people's inherent want/need for NAT that does not need to exist in the IPv6 world. These addresses will always be within the FD00::/8 range, which is part of the FC00::/7 subnet space currently allocate for site local, or "unique local" addresses. The only currently assigned block (for private use) from this space is FD00::/8 however.

 

Lastly your Global Unicast addresses are your real, routable, IPv6 address, with Link Local and Site Local being non-routable. In other words, Global Unicast addresses are the only address that are routable on the open internet and allow for public inter-network communication. The current allocated space for Global Unicast addresses is 2000::/3. Addresses within this space are allocated via regional IP authorities, just link public IPv4 addresses. Allocation of these addresses are restricted to minimum allocations of /48. Anything smaller must come from a local LEC, carrier, or other institution who will "lend" you some of their assigned address space, much like how your current ISP lends you one of their IP's so you can reach the internet.

 

It is important to note that while you can subnet IPv6 addresses below /64, ie. /123 for wan links, this is not recommended for larger network segments as lowering the subnet to blow the 64th bit will cause EUI64, and the variants thereof, to stop functioning, requiring manual assignment of IPv6 addresses or the use of DHCP for IP assignment.

 

The main reason we use NAT in IPv4 is for address preservation. There are far too many devices that need IP addresses and nowhere near enough IPv4 addresses to go around. This lead to the wide-spread adoption of NAT which up until that point, was more-or-less only used in specific situations requiring the obfuscation of IP's and a few other fringe usecases. In the IPv6 world we have more than enough addresses for everyone to have millions, so we just don't need to use NAT in the IPv6 world.

 

NAT in IPv6 does serve an important function in inter-compatibility with IPv4 however in the future when everything is IPv6, this will no longer be a factor.

 

As for the network and host bits of address, they are just used to denote which parts belong to the network (assigned by regional authorities and controlled) and the host bits, which are entirely user controlled. Combined with the subnet mask any computer knows instantly if the address it is trying to reach is part of its own local subnet (network bits matching) or if it is outside its local subnet and requires its default gateway to reach it. Without subnet masks, network bits and host bits, our computers wouldn't have a way to know where they need to send their network traffic.

thanks mate for such a great answer i know that an answer is good to it gives me even more questions then it answerded lol what i still don't understand is this writing FE80::/10 i think it means fe80.0000.0000.0000.0000.0000.0000.0000 but what does the / stand for and i just started so it's normal that i'm totally lost with the new terms like subnet . subnet masks . gateway and what's the diffrence between binary cidr and decimal writings ? 

[Lurick]

11 minutes ago, ilyas001 said:

thanks mate for such a great answer i know that an answer is good to it gives me even more questions then it answerded lol what i still don't understand is this writing FE80::/10 i think it means fe80.0000.0000.0000.0000.0000.0000.0000 but what does the / stand for and i just started so it's normal that i'm totally lost with the new terms like subnet . subnet masks . gateway and what's the diffrence between binary cidr and decimal writings ? 

The /10 is the subnet, so you would have hosts in the range of fe80:0000:0000:0000:0000:0000:0000:0000- febf:ffff:ffff:ffff:ffff:ffff:ffff:ffff

If you're asking about just the slash itself it's just a way to notate the division between the address and the mask, it has no real meaning.

[ilyas001]

16 minutes ago, Lurick said:

The /10 is the subnet, so you would have hosts in the range of fe80:0000:0000:0000:0000:0000:0000:0000- febf:ffff:ffff:ffff:ffff:ffff:ffff:ffff

If you're asking about just the slash itself it's just a way to notate the division between the address and the mask, it has no real meaning.

i don't get it i remember in ipv6  for exemple if it's writed \16 means the first 16 bits is for network  the next is for subnetting and the other are for the host  but what does subnetting do in the first place and wht isn't it on ipv4

and i don't get the \10 or 8 because each one in the ipv6 have 16 bits so the subnet should move by 16 bits each time 10 have no cense to me ? 

and because we use nat for ipv4 why does it have a network or even a host part it's just a number that says what computer on the network needed the information anyway it's gonna be translated by the modem into public adressing so why ? 

 

[leadeater]

1 hour ago, ilyas001 said:

i don't get it i remember in ipv6  for exemple if it's writed \16 means the first 16 bits is for network  the next is for subnetting and the other are for the host  but what does subnetting do in the first place and wht isn't it on ipv4

and i don't get the \10 or 8 because each one in the ipv6 have 16 bits so the subnet should move by 16 bits each time 10 have no cense to me ? 

and because we use nat for ipv4 why does it have a network or even a host part it's just a number that says what computer on the network needed the information anyway it's gonna be translated by the modem into public adressing so why ? 

 

Both IPv4 and IPv6 at the basic address level are fundamentally the same, IPv4 is 32 bits of address space and IPv6 is 128 bits of address space.

 

Both addressing schemes can be translated down and represented in their raw bit level which when learning about IPv4 is common, less so in IPv6 simply due to the length/time it takes to write out all those 1's and 0's.

 

Both support variable length subnet masks so you can have multiple network segments. An example of a use case for this is a subnet per building or a subnet for printers. The backslash '/' is known as the CIDR notation which is a decimal number showing the number of network bits e.g. /16 or 10.1.1.0/16.

 

There are unique things about both IPv4 and IPv6 which @Wombo has already covered some of them.