Annotation of wikisrc/guide/net-intro.mdwn, revision 1.3

1.1       jdf         1: ## Introduction to TCP/IP Networking
                      2: 
                      3: ## Audience
                      4: 
                      5: This section explains various aspects of networking. It is intended to help
                      6: people with little knowledge about networks to get started. It is divided into
                      7: three big parts. We start by giving a general overview of how networking works
                      8: and introduce the basic concepts. Then we go into details for setting up various
                      9: types of networking in the second parts, and the third part of the networking
                     10: section covers a number of *advanced* topics that go beyond the basic operation
                     11: as introduced in the first two sections.
                     12: 
                     13: The reader is assumed to know about basic system administration tasks: how to
                     14: become root, edit files, change permissions, stop processes, etc. See the other
1.3     ! jdf        15: chapters of this NetBSD guide and e.g.*Essential System Administration* by 
        !            16: Aeleen Frisch (1991, O'Reilly & Associates) for further information on this 
        !            17: topic. Besides that, you should know how to handle the utilities we're going to 
        !            18: set up here, i.e. you should know how to use telnet, FTP, ... I will not explain 
        !            19: the basic features of those utilities, please refer to the appropriate 
        !            20: man-pages, the references listed or of course the other parts of this document 
        !            21: instead.
1.1       jdf        22: 
                     23: This introduction to TCP/IP Networking was written with the intention in mind to
                     24: give starters a basic knowledge. If you really want to know what it's all about,
1.3     ! jdf        25: read *TCP/IP Network Administration* by Craig Hunt (1993, O'Reilly & 
        !            26: Associates). This book does not only cover the basics, but goes on and explains 
        !            27: all the concepts, services and how to set them up in detail. It's great, I love 
        !            28: it! :-)
1.1       jdf        29: 
                     30: ## Supported Networking Protocols
                     31: 
                     32: There are several protocol suites supported by NetBSD, most of which were
                     33: inherited from NetBSD's predecessor, 4.4BSD, and subsequently enhanced and
                     34: improved. The first and most important one today is DARPA's Transmission Control
                     35: Protocol/Internet Protocol (TCP/IP). Other protocol suites available in NetBSD
                     36: include the Xerox Network System (XNS) which was only implemented at UCB to
                     37: connect isolated machines to the net, Apple's AppleTalk protocol suite and the
                     38: ISO protocol suite, CCITT X.25 and ARGO TP. They are only used in some special
                     39: applications these days.
                     40: 
                     41: Today, TCP/IP is the most widespread protocol of the ones mentioned above. It is
                     42: implemented on almost every hardware and operating system, and it is also the
                     43: most-used protocol in heterogenous environments. So, if you just want to connect
                     44: your computer running NetBSD to some other machine at home or you want to
                     45: integrate it into your company's or university's network, TCP/IP is the right
                     46: choice. Besides the "old" IP version 4, NetBSD also supports the "new" IP
                     47: version 6 (IPv6) since NetBSD 1.5, thanks to code contributed by the KAME
                     48: project.
                     49: 
                     50: There are other protocol suites such as DECNET, Novell's IPX/SPX or Microsoft's
                     51: NetBIOS, but these are not currently supported by NetBSD. These protocols differ
                     52: from TCP/IP in that they are proprietary, in contrast to the others, which are
                     53: well-defined in several RFCs and other open standards.
                     54: 
                     55: ## Supported Media
                     56: 
                     57: The TCP/IP protocol stack behaves the same regardless of the underlying media
                     58: used, and NetBSD supports a wide range of these, among them are Ethernet
                     59: (10/100/1000MBd), Arcnet, serial line, ATM, FDDI, Fiber Channel, USB, HIPPI,
                     60: FireWire (IEEE 1394), Token Ring, and serial lines.
                     61: 
                     62: ### Serial Line
                     63: 
                     64: There are a couple of reasons for using TCP/IP over a serial line.
                     65: 
                     66:  * If your remote host is only reachable via telephone, you can use a modem to access it.
                     67:  * Many computers have a serial port, and the cable needed is rather cheap.
                     68: 
                     69: The disadvantage of a serial connection is that it's slower than other methods.
                     70: NetBSD can use at most 115200 bit/s, making it a lot slower than e.g. Ethernet's
                     71: minimum 10 Mbit/s and Arcnet's 4 Mbit/s.
                     72: 
                     73: There are two possible protocols to connect a host running NetBSD to another
                     74: host using a serial line (possibly over a phone-line):
                     75: 
                     76:  * Serial Line IP (SLIP)
                     77:  * Point to Point Protocol (PPP)
                     78: 
                     79: The choice here depends on whether you use a dial-up connection through a modem
                     80: or if you use a static connection (null-modem or leased line). If you dial up
                     81: for your IP connection, it's wise to use PPP as it offers some possibilities to
                     82: auto-negotiate IP-addresses and routes, which can be quite painful to do by
                     83: hand. If you want to connect to another machine which is directly connected,
                     84: use SLIP, as this is supported by about every operating system and more easy
                     85: to set up with fixed addresses and routes.
                     86: 
                     87: PPP on a direct connection is a bit difficult to setup, as it's easy to timeout
                     88: the initial handshake; with SLIP, there's no such initial handshake, i.e. you
                     89: start up one side, and when the other site has its first packet, it will send it
                     90: over the line.
                     91: 
                     92: [RFC1331](http://tools.ietf.org/html/rfc1331) ("The Point-to-Point Protocol
                     93: (PPP) for the Transmission of Multi-protocol Datagrams over Point-to-Point
                     94: Links") and [RFC1332](http://tools.ietf.org/html/rfc1332) ("The PPP Internet
                     95: Protocol Control Protocol (IPCP)") describe PPP and TCP/IP over PPP. SLIP is
                     96: defined in [RFC1055](http://tools.ietf.org/html/rfc1055) ("Nonstandard for
                     97: transmission of IP datagrams over serial lines: SLIP").
                     98: 
                     99: ### Ethernet
                    100: 
                    101: Ethernet is the medium commonly used to build local area networks (LANs) of
                    102: interconnected machines within a limited area such as an office, company or
                    103: university campus. Ethernet is based on a bus structure to which many machines
                    104: can connect to, and communication always happens between two nodes at a time.
                    105: When two or more nodes want to talk at the same time, both will restart
                    106: communication after some timeout. The technical term for this is CSMA/CD
                    107: (Carrier Sense w/ Multiple Access and Collision Detection).
                    108: 
                    109: Initially, Ethernet hardware consisted of a thick (yellow) cable that machines
                    110: tapped into using special connectors that poked through the cable's outer
                    111: shielding. The successor of this was called 10base5, which used BNC-type
                    112: connectors for tapping in special T-connectors and terminators on both ends of
                    113: the bus. Today, ethernet is mostly used with twisted pair lines which are used
                    114: in a collapsed bus system that are contained in switches or hubs. The twisted
                    115: pair lines give this type of media its name - 10baseT for 10 Mbit/s networks,
                    116: and 100baseT for 100 MBit/s ones. In switched environments there's also the
                    117: distinction if communication between the node and the switch can happen in half-
                    118: or in full duplex mode.
                    119: 
                    120: ## TCP/IP Address Format
                    121: 
                    122: TCP/IP uses 4-byte (32-bit) addresses in the current implementations (IPv4),
                    123: also called IP-numbers (Internet-Protocol numbers), to address hosts.
                    124: 
                    125: TCP/IP allows any two machines to communicate directly. To permit this all hosts
                    126: on a given network must have a unique IP address. To assure this, IP addresses
                    127: are administrated by one central organisation, the InterNIC. They give certain
                    128: ranges of addresses (network-addresses) directly to sites which want to
                    129: participate in the internet or to internet-providers, which give the addresses
                    130: to their customers.
                    131: 
                    132: If your university or company is connected to the Internet, it has (at least)
                    133: one such network-address for its own use, usually not assigned by the InterNIC
                    134: directly, but rather through an Internet Service Provider (ISP).
                    135: 
                    136: If you just want to run your private network at home, see below on how to
                    137: "build" your own IP addresses. However, if you want to connect your machine to
                    138: the (real :-) Internet, you should get an IP addresses from your local
                    139: network-administrator or -provider.
                    140: 
                    141: IP addresses are usually written in *dotted quad* notation - the four bytes are
                    142: written down in decimal (most significant byte first), separated by dots. For
                    143: example, 132.199.15.99 would be a valid address. Another way to write down
                    144: IP-addresses would be as one 32-bit hex-word, e.g. 0x84c70f63. This is not as
                    145: convenient as the dotted-quad, but quite useful at times, too. (See below!)
                    146: 
                    147: Being assigned a network means nothing else but setting some of the
                    148: above-mentioned 32 address-bits to certain values. These bits that are used for
                    149: identifying the network are called network-bits. The remaining bits can be used
                    150: to address hosts on that network, therefore they are called host-bits. The
                    151: following figure illustrates the separation.
                    152: 
                    153: ![IPv4-addresses are divided into more significant network- and less significant hostbits](/guide/images/ipv4-en-0bits.png)  
                    154: **IPv4-addresses are divided into more significant network- and less significant hostbits**
                    155: 
                    156: In the above example, the network-address is 132.199.0.0 (host-bits are set to 0
                    157: in network-addresses) and the host's address is 15.99 on that network.
                    158: 
                    159: How do you know that the host's address is 16 bit wide? Well, this is assigned
                    160: by the provider from which you get your network-addresses. In the classless
                    161: inter-domain routing (CIDR) used today, host fields are usually between as
                    162: little as 2 to 16 bits wide, and the number of network-bits is written after the
                    163: network address, separated by a `/`, e.g. 132.199.0.0/16 tells that the network
                    164: in question has 16 network-bits. When talking about the *size* of a network,
                    165: it's usual to only talk about it as `/16`, `/24`, etc.
                    166: 
                    167: Before CIDR was used, there used to be four classes of networks. Each one starts
                    168: with a certain bit-pattern identifying it. Here are the four classes:
                    169: 
                    170:  * Class A starts with `0` as most significant bit. The next seven bits of a
                    171:    class A address identify the network, the remaining 24 bit can be used to
                    172:    address hosts. So, within one class A network there can be 2^24 hosts. It's
                    173:    not very likely that you (or your university, or company, or whatever) will
                    174:    get a whole class A address.
                    175: 
                    176:    The CIDR notation for a class A network with its eight network bits is an `/8`.
                    177: 
                    178:  * Class B starts with `10` as most significant bits. The next 14 bits are used
                    179:    for the networks address and the remaining 16 bits can be used to address
                    180:    more than 65000 hosts. Class B addresses are very rarely given out today,
                    181:    they used to be common for companies and universities before IPv4 address
                    182:    space went scarce.
                    183: 
                    184:    The CIDR notation for a class B network with its 16 network bits is an `/16`.
                    185: 
                    186:    Returning to our above example, you can see that 132.199.15.99 (or
                    187:    0x84c70f63, which is more appropriate here!) is on a class B network, as
                    188:    0x84... = **10**00... (base 2).
                    189: 
                    190:    Therefore, the address 132.199.15.99 can be split into an network-address of
                    191:    132.199.0.0 and a host-address of 15.99.
                    192: 
                    193:  * Class C is identified by the MSBs being `110`, allowing only 256 (actually:
                    194:    only 254, see below) hosts on each of the 2^21 possible class C networks.
                    195:    Class C addresses are usually found at (small) companies.
                    196: 
                    197:    The CIDR notation for a class C network with its 24 network bits is an `/24`.
                    198: 
                    199:  * There are also other addresses, starting with `111`. Those are used for
                    200:    special purposes (e. g. multicast-addresses) and are not of interest here.
                    201: 
                    202: Please note that the bits which are used for identifying the network-class are
                    203: part of the network-address.
                    204: 
                    205: When separating host-addresses from network-addresses, the *netmask* comes in
                    206: handy. In this mask, all the network-bits are set to `1`, the host-bits are `0`.
                    207: Thus, putting together IP-address and netmask with a logical AND-function, the
                    208: network-address remains.
                    209: 
                    210: To continue our example, 255.255.0.0 is a possible netmask for 132.199.15.99.
                    211: When applying this mask, the network-address 132.199.0.0 remains.
                    212: 
                    213: For addresses in CIDR notation, the number of network-bits given also says how
                    214: many of the most significant bits of the address must be set to `1` to get the
                    215: netmask for the corresponding network. For classful addressing, every
                    216: network-class has a fixed default netmask assigned:
                    217: 
                    218:  * Class A (/8): default-netmask: 255.0.0.0, first byte of address: 1-127
                    219:  * Class B (/16): default-netmask: 255.255.0.0, first byte of address: 128-191
                    220:  * Class C (/24): default-netmask: 255.255.255.0, first byte of address: 192-223
                    221: 
                    222: Another thing to mention here is the *broadcast-address*. When sending to this
                    223: address, *all* hosts on the corresponding network will receive the message sent.
                    224: The broadcast address is characterized by having all host-bits set to `1`.
                    225: 
                    226: Taking 132.199.15.99 with its netmask 255.255.0.0 again, the broadcast-address
                    227: would result in 132.199.255.255.
                    228: 
                    229: You'll ask now: But what if I want a host's address to be all bits `0` or `1`?
                    230: Well, this doesn't work, as network- and broadcast-address must be present!
                    231: Because of this, a class B (/16) network can contain at most 2^16-2 hosts, a
                    232: class C (/24) network can hold no more than 2^8-2 = 254 hosts.
                    233: 
                    234: Besides all those categories of addresses, there's the special IP-address
                    235: 127.0.0.1 which always refers to the *local* host, i.e. if you talk to 127.0.0.1
                    236: you'll talk to yourself without starting any network-activity. This is sometimes
                    237: useful to use services installed on your own machine or to play around if you
                    238: don't have other hosts to put on your network.
                    239: 
                    240: Let's put together the things we've introduced in this section:
                    241: 
                    242:  * *IP-address* -- 32 bit-address, with network- and host-bits.
                    243:  * *Network-address* -- IP-address with all host bits set to `0`.
                    244:  * *Netmask* -- 32-bit mask with `1` for network- and `0` for host-bits.
                    245:  * *Broadcast* -- IP-address with all host bits set `1`.
                    246:  * *localhost's address* -- The local host's IP address is always 127.0.0.1.
                    247: 
                    248: ## Subnetting and Routing
                    249: 
                    250: After talking so much about netmasks, network-, host- and other addresses, I
                    251: have to admit that this is not the whole truth.
                    252: 
                    253: Imagine the situation at your university, which usually has a class B (/16)
1.2       jdf       254: address, allowing it to have up to 2^16 ~= 65534 hosts on that net. Maybe it
1.1       jdf       255: would be a nice thing to have all those hosts on one single network, but it's
                    256: simply not possible due to limitations in the transport media commonly used
                    257: today.
                    258: 
                    259: For example, when using thinwire ethernet, the maximum length of the cable is
                    260: 185 meters. Even with repeaters in between, which refresh the signals, this is
                    261: not enough to cover all the locations where machines are located. Besides that,
                    262: there is a maximum number of 1024 hosts on one ethernet wire, and you'll lose
                    263: quite a bit of performance if you go to this limit.
                    264: 
                    265: So, are you hosed now? Having an address which allows more than 60000 hosts, but
                    266: being bound to media which allows far less than that limit?
                    267: 
                    268: Well, of course not! :-)
                    269: 
                    270: The idea is to divide the *big* class B net into several smaller networks,
                    271: commonly called sub-networks or simply subnets. Those subnets are only allowed
                    272: to have, say, 254 hosts on them (i.e. you divide one big class B network into
                    273: several class C networks!).
                    274: 
                    275: To do this, you adjust your netmask to have more network- and less host-bits on
                    276: it. This is usually done on a byte-boundary, but you're not forced to do it
                    277: there. So, commonly your netmask will not be 255.255.0.0 as supposed by a class
                    278: B network, but it will be set to 255.255.255.0.
                    279: 
                    280: In CIDR notation, you now write a `/24` instead of the `/16` to show that 24
                    281: bits of the address are used for identifying the network and subnet, instead of
                    282: the 16 that were used before.
                    283: 
                    284: This gives you one additional network-byte to assign to each (physical!)
                    285: network. All the 254 hosts on that subnet can now talk directly to each other,
                    286: and you can build 256 such class C nets. This should fit your needs.
                    287: 
                    288: To explain this better, let's continue our above example. Say our host
                    289: 132.199.15.99 (I'll call him `dusk` from now; we'll talk about assigning
                    290: hostnames later) has a netmask of 255.255.255.0 and thus is on the subnet
                    291: 132.199.15.0/24. Let's furthermore introduce some more hosts so we have
                    292: something to play around with, see the next figure.
                    293: 
                    294: ![Our demo-network](/guide/images/net-pic1.gif)  
                    295: **Our demo-network**
                    296: 
                    297: In the above network, `dusk` can talk directly to `dawn`, as they are both on
                    298: the same subnet. (There are other hosts attached to the 132.199.15.0/24-subnet
                    299: but they are not of importance for us now)
                    300: 
                    301: But what if `dusk` wants to talk to a host on another subnet?
                    302: 
                    303: Well, the traffic will then go through one or more gateways (routers), which are
                    304: attached to two subnets. Because of this, a router always has two different
                    305: addresses, one for each of the subnets it is on. The router is functionally
                    306: transparent, i.e. you don't have to address it to reach hosts on the *other*
                    307: side. Instead, you address that host directly and the packets will be routed to
                    308: it correctly.
                    309: 
                    310: Example. Let's say `dusk` wants to get some files from the local ftp-server. As
                    311: `dusk` can't reach `ftp` directly (because it's on a different subnet), all its
                    312: packets will be forwarded to its "defaultrouter" `rzi` (132.199.15.1), which
                    313: knows where to forward the packets.
                    314: 
                    315: `Dusk` knows the address of its defaultrouter in its network (`rzi`,
                    316: 132.199.15.1), and it will forward any packets to it which are not on the same
                    317: subnet, i.e. it will forward all IP-packets in which the third address-byte
                    318: isn't 15.
                    319: 
                    320: The (default)router then gives the packets to the appropriate host, as it's also
                    321: on the FTP-server's network.
                    322: 
                    323: In this example, *all* packets are forwarded to the 132.199.1.0/24-network,
                    324: simply because it's the network's backbone, the most important part of the
                    325: network, which carries all the traffic that passes between several subnets.
                    326: Almost all other networks besides 132.199.15.0/24 are attached to the backbone
                    327: in a similar manner.
                    328: 
                    329: But what if we had hooked up another subnet to 132.199.15.0/24 instead of
                    330: 132.199.1.0/24? Maybe something the situation displayed in the next figure.
                    331: 
                    332: ![Attaching one subnet to another one](/guide/images/net-pic2.gif)  
                    333: **Attaching one subnet to another one**
                    334: 
                    335: When we now want to reach a host which is located in the 132.199.16.0/24-subnet
                    336: from `dusk`, it won't work routing it to `rzi`, but you'll have to send it
                    337: directly to `route2` (132.199.15.2). `Dusk` will have to know to forward those
                    338: packets to `route2` and send all the others to `rzi`.
                    339: 
                    340: When configuring `dusk`, you tell it to forward all packets for the
                    341: 132.199.16.0/24-subnet to `route2`, and all others to `rzi`. Instead of
                    342: specifying this default as 132.199.1.0/24, 132.199.2.0/24, etc., 0.0.0.0 can be
                    343: used to set the default-route.
                    344: 
                    345: Returning to our demo network, there's a similar problem when `dawn` wants to
                    346: send to `noon`, which is connected to `dusk` via a serial line running. When
                    347: looking at the IP-addresses, `noon` seems to be attached to the
                    348: 132.199.15.0-network, but it isn't really. Instead, `dusk` is used as gateway,
                    349: and `dawn` will have to send its packets to `dusk`, which will forward them to
                    350: `noon` then. The way `dusk` is forced into accepting packets that aren't
                    351: destined at it but for a different host (`noon`) instead is called *proxy arp*.
                    352: 
                    353: The same goes when hosts from other subnets want to send to `noon`. They have to
                    354: send their packets to `dusk` (possibly routed via `rzi`),
                    355: 
                    356: ## Name Service Concepts
                    357: 
                    358: In the previous sections, when we talked about hosts, we referred to them by
                    359: their IP-addresses. This was necessary to introduce the different kinds of
                    360: addresses. When talking about hosts in general, it's more convenient to give
                    361: them *names*, as we did when talking about routing.
                    362: 
                    363: Most applications don't care whether you give them an IP address or a hostname.
                    364: However, they'll use IP addresses internally, and there are several methods for
                    365: them to map hostnames to IP addresses, each one with its own way of
                    366: configuration. In this section we'll introduce the idea behind each method, in
                    367: the next chapter, we'll talk about the configuration-part.
                    368: 
                    369: The mapping from hostnames (and domainnames) to IP-addresses is done by a piece
                    370: of software called the *resolver*. This is not an extra service, but some
                    371: library routines which are linked to every application using networking-calls.
                    372: The resolver will then try to resolve (hence the name ;-) the hostnames you give
                    373: into IP addresses. See [RFC1034](http://tools.ietf.org/html/rfc1034) ("Domain
                    374: names - concepts and facilities") and
                    375: [RFC1035](http://tools.ietf.org/html/rfc1035) ("Domain names - implementation
                    376: and specification") for details on the resolver.
                    377: 
                    378: Hostnames are usually up to 256 characters long, and contain letters, numbers
                    379: and dashes (`-`); case is ignored.
                    380: 
                    381: Just as with networks and subnets, it's possible (and desirable) to group hosts
                    382: into domains and subdomains. When getting your network-address, you usually also
                    383: obtain a domainname by your provider. As with subnets, it's up to you to
                    384: introduce subdomains. Other as with IP-addresses, (sub)domains are not directly
                    385: related to (sub)nets; for example, one domain can contain hosts from several
                    386: subnets.
                    387: 
                    388: Our demo-network shows this: Both subnets 132.199.1.0/24 and 132.199.15.0/24
                    389: (and others) are part of the subdomain `rz.uni-regensburg.de`. The domain the
                    390: University of Regensburg got from its IP-provider is `uni-regensburg.de` (`.de`
                    391: is for Deutschland, Germany), the subdomain `rz` is for Rechenzentrum, computing
                    392: center.
                    393: 
                    394: Hostnames, subdomain- and domainnames are separated by dots (`.`). It's also
                    395: possible to use more than one stage of subdomains, although this is not very
                    396: common. An example would be `fox_in.socs.uts.edu.au`.
                    397: 
                    398: A hostname which includes the (sub)domain is also called a fully qualified
                    399: domain name (FQDN). For example, the IP-address 132.199.15.99 belongs to the
                    400: host with the FQDN `dusk.rz.uni-regensburg.de`.
                    401: 
                    402: Further above I told you that the IP-address 127.0.0.1 always belongs to the
                    403: local host, regardless what's the `real` IP-address of the host. Therefore,
                    404: 127.0.0.1 is always mapped to the name `localhost`.
                    405: 
                    406: The three different ways to translate hostnames into IP addresses are:
                    407: `/etc/hosts`, the Domain Name Service (DNS) and the Network Information Service
                    408: (NIS).
                    409: 
1.2       jdf       410: ### /etc/hosts
1.1       jdf       411: 
                    412: The first and simplest way to translate hostnames into IP-addresses is by using
                    413: a table telling which IP address belongs to which hostname(s). This table is
                    414: stored in the file `/etc/hosts` and has the following format:
                    415: 
                    416:     IP-address        hostname [nickname [...]]
                    417: 
                    418: Lines starting with a hash mark (`#`) are treated as comments. The other lines
                    419: contain one IP-address and the corresponding hostname(s).
                    420: 
                    421: It's not possible for a hostname to belong to several IP addresses, even if I
                    422: made you think so when talking about routing. `rzi` for example has really two
                    423: distinct names for each of its two addresses: `rzi` and `rzia` (but please don't
                    424: ask me which name belongs to which address!).
                    425: 
                    426: Giving a host several nicknames can be convenient if you want to specify your
                    427: favorite host providing a special service with that name, as is commonly done
                    428: with FTP-servers. The first (leftmost) name is usually the real (canonical) name
                    429: of the host.
                    430: 
                    431: Besides giving nicknames, it's also convenient to give a host's full name
                    432: (including domain) as its canonical name, and using only its hostname (without
                    433: domain) as a nickname.
                    434: 
                    435: *Important:* There *must* be an entry mapping localhost to 127.0.0.1 in
                    436: `/etc/hosts`!
                    437: 
                    438: ### Domain Name Service (DNS)
                    439: 
                    440: `/etc/hosts` bears an inherent problem, especially in big networks: when one
                    441: host is added or one host's address changes, all the `/etc/hosts` files on all
                    442: machines have to be changed! This is not only time-consuming, it's also very
                    443: likely that there will be some errors and inconsistencies, leading to problems.
                    444: 
                    445: Another approach is to hold only one hostnames-table (-database) for a network,
                    446: and make all the clients query that *nameserver*. Updates will be made only on
                    447: the nameserver.
                    448: 
                    449: This is the basic idea behind the Domain Name Service (DNS).
                    450: 
                    451: Usually, there's one nameserver for each domain (hence DNS), and every host
                    452: (client) in that domain knows which domain it is in and which nameserver to
                    453: query for its domain.
                    454: 
                    455: When the DNS gets a query about a host which is not in its domain, it will
                    456: forward the query to a DNS which is either the DNS of the domain in question or
                    457: knows which DNS to ask for the specified domain. If the DNS forwarded the query
                    458: doesn't know how to handle it, it will forward that query again to a DNS one
                    459: step higher. This is not ad infinitum, there are several *root*-servers, which
                    460: know about any domain.
                    461: 
                    462: See [[the separate article|guide/dns]] for details on DNS.
                    463: 
                    464: ### Network Information Service (NIS/YP)
                    465: 
                    466: Yellow Pages (YP) was invented by Sun Microsystems. The name has been changed
                    467: into Network Information Service (NIS) because YP was already a trademark of the
                    468: British telecom. So, when I'm talking about NIS you'll know what I mean. ;-)
                    469: 
                    470: There are quite some configuration files on a Unix-system, and often it's
                    471: desired to maintain only one set of those files for a couple of hosts. Those
                    472: hosts are grouped together in a NIS-domain (which has *nothing* to do with the
                    473: domains built by using DNS!) and are usually contained in one workstation
                    474: cluster.
                    475: 
                    476: Examples for the config-files shared among those hosts are `/etc/passwd`,
                    477: `/etc/group` and - last but not least - `/etc/hosts`.
                    478: 
                    479: So, you can "abuse" NIS for getting a unique name-to-address-translation on all
                    480: hosts throughout one (NIS-)domain.
                    481: 
                    482: There's only one drawback, which prevents NIS from actually being used for that
                    483: translation: In contrast to the DNS, NIS provides no way to resolve hostnames
                    484: which are not in the hosts-table. There's no hosts "one level up" which the
                    485: NIS-server can query, and so the translation will fail! Suns NIS+ takes measures
                    486: against that problem, but as NIS+ is only available on Solaris-systems, this is
                    487: of little use for us now.
                    488: 
                    489: Don't get me wrong: NIS is a fine thing for managing e.g. user-information
                    490: (`/etc/passwd`, ...) in workstation-clusters, it's simply not too useful for
                    491: resolving hostnames.
                    492: 
                    493: ### Other
                    494: 
                    495: The name resolving methods described above are what's used commonly today to
                    496: resolve hostnames into IP addresses, but they aren't the only ones. Basically,
                    497: every database mechanism would do, but none is implemented in NetBSD. Let's have
                    498: a quick look what you may encounter.
                    499: 
                    500: With NIS lacking hierarchy in data structures, NIS+ is intended to help out in
                    501: that field. Tables can be setup in a way so that if a query cannot be answered
                    502: by a domain's server, there can be another domain "above" that might be able to
                    503: do so. E.g. you could choose to have a domain that lists all the hosts (users,
                    504: groups, ...) that are valid in the whole company, one that defines the same
                    505: for each division, etc. NIS+ is not used a lot today, even Sun went back to
                    506: ship back NIS by default.
                    507: 
                    508: Last century, the X.500 standard was designed to accommodate both simple
                    509: databases like `/etc/hosts` as well as complex, hierarchical systems as can be
                    510: found e.g. in DNS today. X.500 wasn't really a success, mostly due to the fact
                    511: that it tried to do too much at the same time. A cut-down version is available
                    512: today as the Lightweight Directory Access Protocol (LDAP), which is becoming
                    513: popular in the last years to manage data like users but also hosts and others in
                    514: small to medium sized organisations.
                    515: 
                    516: ## Next generation Internet protocol - IPv6
                    517: 
                    518: ### The Future of the Internet
                    519: 
                    520: According to experts, the Internet as we know it will face a serious problem in
                    521: a few years. Due to its rapid growth and the limitations in its design, there
                    522: will be a point at which no more free addresses are available for connecting new
                    523: hosts. At that point, no more new web servers can be set up, no more users can
                    524: sign up for accounts at ISPs, no more new machines can be setup to access the
                    525: web or participate in online games - some people may call this a serious
                    526: problem.
                    527: 
                    528: Several approaches have been made to solve the problem. A very popular one is to
                    529: not assign a worldwide unique address to every user's machine, but rather to
                    530: assign them *private* addresses, and hide several machines behind one official,
                    531: globally unique address. This approach is called *Network Address Translation*
                    532: (NAT, also known as IP Masquerading). It has problems, as the machines hidden
                    533: behind the global address can't be addressed, and as a result of this, opening
                    534: connections to them - which is used in online gaming, peer to peer networking,
                    535: etc. - is not possible. For a more in-depth discussion of the drawbacks of NAT,
                    536: see \[[[RFC3027|guide/index#bilbiography]]\] ("Protocol Complications with the
                    537: IP Network Address Translator").
                    538: 
                    539: A different approach to the problem of internet addresses getting scarce is to
                    540: abandon the old Internet protocol with its limited addressing capabilities, and
                    541: use a new protocol that does not have these limitations. The protocol - or
                    542: actually, a set of protocols - used by machines connected to form today's
                    543: Internet is know as the TCP/IP (Transmission Control Protocol, Internet
                    544: Protocol) suite, and version 4 currently in use has all the problems described
                    545: above. Switching to a different protocol version that does not have these
                    546: problems of course requires for a 'better' version to be available, which
                    547: actually is. Version 6 of the Internet Protocol (IPv6) does fulfill any possible
                    548: future demands on address space, and also addresses further features such as
                    549: privacy, encryption, and better support of mobile computing.
                    550: 
                    551: Assuming a basic understanding of how today's IPv4 works, this text is intended
                    552: as an introduction to the IPv6 protocol. The changes in address formats and name
                    553: resolution are covered. With the background given here, the next sections will
                    554: show how to use IPv6 even if your ISP doesn't offer it by using a simple yet
                    555: efficient transition mechanism called 6to4. The goal is to get online with IPv6,
                    556: giving an example configuration for NetBSD.
                    557: 
                    558: ### What good is IPv6?
                    559: 
                    560: When telling people to migrate from IPv4 to IPv6, the question you usually hear
                    561: is *why?*. There are actually a few good reasons to move to the new version:
                    562: 
                    563:  * Bigger address space
                    564:  * Support for mobile devices
                    565:  * Built-in security
                    566: 
                    567: #### Bigger Address Space
                    568: 
                    569: The bigger address space that IPv6 offers is the most obvious enhancement it has
                    570: over IPv4. While today's internet architecture is based on 32-bit wide
                    571: addresses, the new version has 128 bit available for addressing. Thanks to the
                    572: enlarged address space, work-arounds like NAT don't have to be used any more.
                    573: This allows full, unconstrained IP connectivity for today's IP based machines as
                    574: well as upcoming mobile devices like PDAs and cell phones will benefit from full
                    575: IP access through GPRS and UMTS.
                    576: 
                    577: #### Mobility
                    578: 
                    579: When mentioning mobile devices and IP, another important point to note is that
                    580: some special protocol is needed to support mobility, and implementing this
                    581: protocol - called *Mobile IP* - is one of the requirements for every IPv6 stack.
                    582: Thus, if you have IPv6 going, you have support for roaming between different
                    583: networks, with everyone being updated when you leave one network and enter the
                    584: other one. Support for roaming is possible with IPv4 too, but there are a number
                    585: of hoops that need to be jumped in order to get things working. With IPv6,
                    586: there's no need for this, as support for mobility was one of the design
                    587: requirements for IPv6. See [RFC3024](http://tools.ietf.org/html/rfc3024)
                    588: ("Reverse Tunneling for Mobile IP") for some more information on the issues that
                    589: need to be addressed with Mobile IP on IPv4.
                    590: 
                    591: #### Security
                    592: 
                    593: Besides support for mobility, security was another requirement for the successor
                    594: to today's Internet Protocol version. As a result, IPv6 protocol stacks are
                    595: required to include IPsec. IPsec allows authentication, encryption and
                    596: compression of any IP traffic. Unlike application level protocols like SSL or
                    597: SSH, all IP traffic between two nodes can be handled, without adjusting any
                    598: applications. The benefit of this is that all applications on a machine can
                    599: benefit from encryption and authentication, and that policies can be set on a
                    600: per-host (or even per-network) base, not per application/service. An
                    601: introduction to IPsec with a roadmap to the documentation can be found in
                    602: [RFC2411](http://tools.ietf.org/html/rfc2411) ("IP Security Document Roadmap"),
                    603: the core protocol is described in [RFC2401](http://tools.ietf.org/html/rfc2401)
                    604: ("Security Architecture for the Internet Protocol").  
                    605: 
                    606: ### Changes to IPv4
                    607: 
                    608: After giving a brief overview of all the important features of IPv6, we'll go
                    609: into the details of the basics of IPv6 here. A brief understanding of how IPv4
                    610: works is assumed, and the changes in IPv6 will be highlighted. Starting with
                    611: IPv6 addresses and how they're split up we'll go into the various types of
                    612: addresses there are, what became of broadcasts, then after discussing the IP
                    613: layer go into changes for name resolving and what's new in DNS for IPv6.
                    614: 
                    615: #### Addressing
                    616: 
                    617: An IPv4 address is a 32 bit value, that's usually written in *dotted quad*
                    618: representation, where each *quad* represents a byte value between 0 and 255, for
                    619: example:
                    620: 
                    621:     127.0.0.1
                    622: 
1.2       jdf       623: This allows a theoretical number of 2^32 or ~4 billion hosts to be connected on
1.1       jdf       624: the internet today. Due to grouping, not all addresses are available today.
                    625: 
                    626: IPv6 addresses use 128 bit, which results in 2^128 theoretically addressable
                    627: hosts. This allows for a Really Big number of machines to addressed, and it sure
                    628: fits all of today's requirements plus all those nifty PDAs and cell phones with
                    629: IP phones in the near future without any sweat. When writing IPv6 addresses,
                    630: they are usually divided into groups of 16 bits written as four hex digits, and
                    631: the groups are separated by colons. An example is:
                    632: 
                    633:     fe80::2a0:d2ff:fea5:e9f5
                    634: 
                    635: This shows a special thing - a number of consecutive zeros can be abbreviated by
                    636: a single `::` once in the IPv6 address. The above address is thus equivalent to
                    637: `fe80:0:00:000:2a0:d2ff:fea5:e9f5` - leading zeros within groups can be omitted,
                    638: and only one `::` can be used in an IPv6 address.
                    639: 
                    640: To make addresses manageable, they are split in two parts, which are the bits
                    641: identifying the network a machine is on, and the bits that identify a machine on
                    642: a (sub)network. The bits are known as netbits and hostbits, and in both IPv4 and
                    643: IPv6, the netbits are the *left*, most significant bits of an IP address, and
                    644: the host bits are the *right*, least significant bits, as shown in the following
                    645: figure:
                    646: 
                    647: ![IPv6-addresses are divided into more significant network- and less significant hostbits, too](/guide/images/ipv6-en-0bits.gif)  
                    648: **IPv6-addresses are divided into more significant network- and less significant hostbits, too**
                    649: 
                    650: In IPv4, the border is drawn with the aid of the netmask, which can be used to
                    651: mask all net/host bits. Typical examples are 255.255.0.0 that uses 16 bit for
                    652: addressing the network, and 16 bit for the machine, or 255.255.255.0 which takes
                    653: another 8 bit to allow addressing 256 subnets on e.g. a class B net.
                    654: 
                    655: When addressing switched from classful addressing to CIDR routing, the borders
                    656: between net and host bits stopped being on 8 bit boundaries, and as a result the
                    657: netmasks started looking ugly and not really manageable. As a replacement, the
                    658: number of network bits is used for a given address, to denote the border, e.g.
                    659: 
                    660:     10.0.0.0/24
                    661: 
                    662: is the same as a netmask of 255.255.255.0 (24 1-bits). The same scheme is used
                    663: in IPv6:
                    664: 
                    665:     2001:638:a01:2::/64
                    666: 
                    667: tells us that the address used here has the first (leftmost) 64 bits used as the
                    668: network address, and the last (rightmost) 64 bits are used to identify the
                    669: machine on the network. The network bits are commonly referred to as (network)
                    670: *prefix*, and the *prefixlen* here would be 64 bits.
                    671: 
                    672: Common addressing schemes found in IPv4 are the (old) class B and class C nets.
                    673: With a class C network (/24), you get 24 bits assigned by your provider, and it
                    674: leaves 8 bits to be assigned by you. If you want to add any subnetting to that,
                    675: you end up with *uneven* netmasks that are a bit nifty to deal with. Easier for
                    676: such cases are class B networks (/16), which only have 16 bits assigned by the
                    677: provider, and that allow subnetting, i.e. splitting of the rightmost bits into
                    678: two parts. One to address the on-site subnet, and one to address the hosts on
                    679: that subnet. Usually, this is done on byte (8 bit) boundaries. Using a netmask
                    680: of 255.255.255.0 (or a /24 prefix) allows flexible management even of bigger
                    681: networks here. Of course there is the upper limit of 254 machines per subnet,
                    682: and 256 subnets.
                    683: 
                    684: With 128 bits available for addressing in IPv6, the scheme commonly used is the
                    685: same, only the fields are wider. Providers usually assign /48 networks, which
                    686: leaves 16 bits for a subnetting and 64 hostbits.
                    687: 
1.2       jdf       688: ![IPv6-addresses have a similar structure to class B addresses](/guide/images/ipv6-en-6adrformats.gif)  
1.1       jdf       689: **IPv6-addresses have a similar structure to class B addresses**
                    690: 
                    691: Now while the space for network and subnets here is pretty much ok, using 64
                    692: bits for addressing hosts seems like a waste. It's unlikely that you will want
                    693: to have several billion hosts on a single subnet, so what is the idea behind
                    694: this?
                    695: 
                    696: The idea behind fixed width 64 bit wide host identifiers is that they aren't
                    697: assigned manually as it's usually done for IPv4 nowadays. Instead, IPv6 host
                    698: addresses are recommended (not mandatory!) to be built from so-called EUI64
                    699: addresses. EUI64 addresses are - as the name says - 64 bit wide, and derived
                    700: from MAC addresses of the underlying network interface. E.g. for ethernet, the 6
                    701: byte (48 bit) MAC address is usually filled with the hex bits `fffe` in the
                    702: middle and a bit is set to mark the address as unique (which is true for
                    703: Ethernet), e.g. the MAC address
                    704: 
                    705:     01:23:45:67:89:ab
                    706: 
                    707: results in the EUI64 address
                    708: 
                    709:     03:23:45:ff:fe:67:89:ab
                    710: 
                    711: which again gives the host bits for the IPv6 address as
                    712: 
                    713:     ::0323:45ff:fe67:89ab
                    714: 
                    715: These host bits can now be used to automatically assign IPv6 addresses to hosts,
                    716: which supports autoconfiguration of IPv6 hosts - all that's needed to get a
                    717: complete IPv6 address is the first (net/subnet) bits, and IPv6 also offers a
                    718: solution to assign them automatically.
                    719: 
                    720: When on a network of machines speaking IP, there's usually one router which acts
                    721: as the gateway to outside networks. In IPv6 land, this router will send *router
                    722: advertisement* information, which clients are expected to either receive during
                    723: operation or to solicit upon system startup. The router advertisement
                    724: information includes data on the router's address, and which address prefix it
                    725: routes. With this information and the host-generated EUI64 address, an IPv6-host
                    726: can calculate its IP address, and there is no need for manual address
                    727: assignment. Of course routers still need some configuration.
                    728: 
                    729: The router advertisement information they create are part of the Neighbor
                    730: Discovery Protocol (NDP, see [RFC2461](http://tools.ietf.org/html/rfc2461)
                    731: ("Neighbor Discovery for IP Version 6 (IPv6)")), which is the successor to
                    732: IPv4's ARP protocol. In contrast to ARP, NDP does not only do lookup of IPv6
                    733: addresses for MAC addresses (the neighbor solicitation/advertisement part), but
                    734: also does a similar service for routers and the prefixes they serve, which is
                    735: used for autoconfiguration of IPv6 hosts as described in the previous paragraph.
                    736: 
                    737: #### Multiple Addresses
                    738: 
                    739: In IPv4, a host usually has one IP address per network interface or even per
                    740: machine if the IP stack supports it. Only very rare applications like web
                    741: servers result in machines having more than one IP address. In IPv6, this is
                    742: different. For each interface, there is not only a globally unique IP address,
                    743: but there are two other addresses that are of interest: The link local address,
                    744: and the site local address. The link local address has a prefix of fe80::/64,
                    745: and the host bits are built from the interface's EUI64 address. The link local
                    746: address is used for contacting hosts and routers on the same network only, the
                    747: addresses are not visible or reachable from different subnets. If wanted,
                    748: there's the choice of either using global addresses (as assigned by a provider),
                    749: or using site local addresses. Site local addresses are assigned the network
                    750: address fec0::/10, and subnets and hosts can be addressed just as for
                    751: provider-assigned networks. The only difference is, that the addresses will not
                    752: be visible to outside machines, as these are on a different network, and their
                    753: *site local* addresses are in a different physical net (if assigned at all). As
                    754: with the 10/8 network in IPv4, site local addresses can be used, but don't have
                    755: to. For IPv6 it's most common to have hosts assigned a link-local and a global
                    756: IP address. Site local addresses are rather uncommon today, and are no
                    757: substitute for globally unique addresses if global connectivity is required.
                    758: 
                    759: #### Multicasting
                    760: 
                    761: In IP land, there are three ways to talk to a host: unicast, broadcast and
                    762: multicast. The most common one is by talking to it directly, using its unicast
                    763: address. In IPv4, the unicast address is the *normal* IP address assigned to a
                    764: single host, with all address bits assigned. The broadcast address used to
                    765: address all hosts in the same IP subnet has the network bits set to the network
                    766: address, and all host bits set to `1` (which can be easily done using the
                    767: netmask and some bit operations). Multicast addresses are used to reach a number
                    768: of hosts in the same multicast group, which can be machines spread over the
                    769: whole internet. Machines must join multicast groups explicitly to participate,
                    770: and there are special IPv4 addresses used for multicast addresses, allocated
                    771: from the 224/8 subnet. Multicast isn't used very much in IPv4, and only few
                    772: applications like the MBone audio and video broadcast utilities use it.
                    773: 
                    774: In IPv6, unicast addresses are used the same as in IPv4, no surprise there - all
                    775: the network and host bits are assigned to identify the target network and
                    776: machine. Broadcasts are no longer available in IPv6 in the way they were in
                    777: IPv4, this is where multicasting comes into play. Addresses in the ff::/8
                    778: network are reserved for multicast applications, and there are two special
                    779: multicast addresses that supersede the broadcast addresses from IPv4. One is the
                    780: *all routers* multicast address, the others is for *all hosts*. The addresses
                    781: are specific to the subnet, i.e. a router connected to two different subnets can
                    782: address all hosts/routers on any of the subnets it's connected to. Addresses
                    783: here are:
                    784: 
                    785:  * ff0*`X`*::1 for all hosts and
                    786:  * ff0*`X`*::2 for all routers,
                    787: 
                    788: where `X` is the scope ID of the link here, identifying the network. Usually
                    789: this starts from `1` for the *node local* scope, `2` for the first link, etc.
                    790: Note that it's perfectly ok for two network interfaces to be attached to one
                    791: link, thus resulting in double bandwidth:
                    792: 
                    793: ![Several interfaces attached to a link result in only one scope ID for the link](/guide/images/ipv6-en-4scope.gif)  
                    794: **Several interfaces attached to a link result in only one scope ID for the link**
                    795: 
                    796: One use of the *all hosts* multicast is in the neighbor solicitation code of
                    797: NDP, where any machine that wants to communicate with another machine sends out
                    798: a request to the *all hosts* group, and the machine in question is expected to
                    799: respond.
                    800: 
                    801: #### Name Resolving in IPv6
                    802: 
                    803: After talking a lot about addressing in IPv6, anyone still here will hope that
                    804: there's a proper way to abstract all these long & ugly IPv6 addresses with some
                    805: nice hostnames as one can do in IPv4, and of course there is.
                    806: 
                    807: Hostname to IP address resolving in IPv4 is usually done in one of three ways:
                    808: using a simple table in `/etc/hosts`, by using the Network Information Service
                    809: (NIS, formerly YP) or via the Domain Name System (DNS).
                    810: 
                    811: As of this writing, NIS/NIS+ over IPv6 is currently only available on Solaris 8,
                    812: for both database contents and transport, using a RPC extension.
                    813: 
                    814: Having a simple address<-\>name map like `/etc/hosts` is supported in all IPv6
                    815: stacks. With the KAME implementation used in NetBSD, `/etc/hosts` contains IPv6
                    816: addresses as well as IPv4 addresses. A simple example is the `localhost` entry
                    817: in the default NetBSD installation:
                    818: 
                    819:     127.0.0.1               localhost
                    820:     ::1                     localhost
                    821: 
                    822: For DNS, there are no fundamentally new concepts. IPv6 name resolving is done
                    823: with AAAA records that - as the name implies - point to an entity that's four
                    824: times the size of an A record. The AAAA record takes a hostname on the left
                    825: side, just as A does, and on the right side there's an IPv6 address, e.g.
                    826: 
                    827:     noon            IN      AAAA    3ffe:400:430:2:240:95ff:fe40:4385
                    828: 
                    829: For reverse resolving, IPv4 uses the in-addr.arpa zone, and below that it writes
                    830: the bytes (in decimal) in reversed order, i.e. more significant bytes are more
                    831: right. For IPv6 this is similar, only that hex digits representing 4 bits are
                    832: used instead of decimal numbers, and the resource records are also under a
                    833: different domain, ip6.int.
                    834: 
                    835: So to have the reverse resolving for the above host, you would put into your
                    836: `/etc/named.conf` something like:
                    837: 
                    838:     zone "0.3.4.0.0.0.4.0.e.f.f.3.IP6.INT" {
                    839:           type master;
                    840:           file "db.reverse";
                    841:     };
                    842: 
                    843: and in the zone file db.reverse you put (besides the usual records like SOA and
                    844: NS):
                    845: 
                    846:     5.8.3.4.0.4.e.f.f.f.5.9.0.4.2.0.2.0.0.0 IN   PTR   noon.ipv6.example.com.
                    847: 
                    848: The address is reversed here, and written down one hex digit after the other,
                    849: starting with the least significant (rightmost) one, separating the hex digits
                    850: with dots, as usual in zone files.
                    851: 
                    852: One thing to note when setting up DNS for IPv6 is to take care of the DNS
                    853: software version in use. BIND 8.x does understand AAAA records, but it does not
                    854: offer name resolving via IPv6. You need BIND 9.x for that. Beyond that, BIND 9.x
                    855: supports a number of resource records that are currently being discussed but not
                    856: officially introduced yet. The most noticeable one here is the A6 record which
                    857: allows easier provider/prefix changing.
                    858: 
                    859: To sum up, this section talked about the technical differences between IPv4 and
                    860: IPv6 for addressing and name resolving. Some details like IP header options, QoS
                    861: and flows were deliberately left out to not make this document more complex than
                    862: necessary.
                    863: 

CVSweb for NetBSD wikisrc <wikimaster@NetBSD.org> software: FreeBSD-CVSweb