This is really a matter of convention rather than a technical consideration these days, as classful addresses were phased out in 1993.
So the notion of "classful" vs "classless" subnetting doesn't really come up any more, except as a convienience. People will generally only use "class b" as easy way to convey a 16bit network address. No one thinks of classes when planning network address space.
Trying to maintain the notion of classful addressing, and introducing a subnet id to distinguish between the classful and non-classful parts of the address is backward, and as you have discovered, confusing.
There aren't really three parts to the address, there are only two. The subnet id is not configured anywhere. There are two parts to an IP address, the network and the host.
Your example shows a 27bit network address with a 5bit host address.
No, switches are not just simplified routers. Although many devices combine functions of both routing and switching, the two functions are distinct. Switches create networks, routers connect distinct networks together. Switches operate using only MAC-addresses, while routers also use IP addresses. Switches have many ports in the same subnet, while routers can only have one port per subnet. Devices that combine these functions still maintain a distinction between routing and switching, as they will separate 'switched' ports from 'routed' ports due to the requirements of the different functions and hardware.
To show some of the nuances: professional grade routers will perform internal switching functions (such as CEF), and some 'layer 3' switches Route (if they are layer-3 switches), but the operations are distinct.
It is useful for anyone interested in networking to become familiar with the OSI Model, which describes how computers and other devices talk to each other. The OSI Model breaks communication into seven layers, including the Application layer, which is the program you, the human, are working with, the Network layer or Layer-3, which deals with IP addresses and Routing, and the Data Link layer, or Layer-2, which deals with physical hardware addresses. Also, when you hear terms like TCP, ports, sockets, sessions, etc., these are also represented by the OSI model, so it's useful to learn.
Routers operate with IP addresses at the OSI Layer-3 or Network Layer and Switches operate with MAC addresses at the OSI Layer 2 or Data Link Layer.
Some devices, such as consumer wifi-routers, combine both a switch and a router in the same device (for instance a wifi router with 5 extra ports on the back). Other devices, such as cable or DSL modems, still function as a router because they route packets between your local network, and the remote network, and perform layer-3 functions such as NAT.
The operations of Routing and Switching are distinct however, and operate on different layers of encapsulation, different OSI layers, and with different addresses, as described below:
In detail:
Switching
We will start with a Layer-2 Switch. This is your standard switch, that includes all un-managed switches, and even managed switches that only operate at the Data-Link layer. This switch receives frames and compares them to a MAC-Address-Table (which is distinct from the ARP table in that it has MAC Addresses and Ports, rather than MACs and IPs). It uses this information to forward frames either to one, many, or all ports depending on whether it is a unicast, multicast, or broadcast frame. If it is a unicast frame, but it does not know which port to send it to, it floods it to ALL ports, except the port the switch received the frame on. Of course there is more detail with CAM tables and VLANS, but in short: Switching moves frames based on hardware addresses, switching can only move frames within the same subnet.
Routing
A router routes between subnets. In fact, on a router, you cannot put multiple routed interfaces on the same subnet - because a routers function is to move packets between layer-3 subnets. The router thus receives frames, containing packets off a port. If frame's destination (mac-address) matches the router (either as unicast or broadcast), the router will then look at the IP-packet contained inside the frame, and make a routing decision based on the IP and subnets. Once the routing decision is made, a switching decision is made to determine which port and layer-2 destination to send the packet to, so it can be re-encapsulated at Layer-2. For more information about the switching conducted by the router, see Cisco Express Forwarding, for example.
How the computer gets through the gateway:
So how does the computer reach its gateway? The computer itself has a route to its gateway. It knows that "if an IP is not on my subnet, I need to send the packet to this gateway address to get out," Thus, since all devices process top-down through the OSI-model, the computer will do an internal routing lookup first (layer-3) to decide wither a device is within its subnet. If not, it will arp (layer-2) to find the mac-address of the default gateway, and will encapsulate the IP packet for the remote host within a frame addressed to the router. The frame will be switched across the network using layer-2 headers and MAC addresses, within the same subnet, until it reaches the router, where its layer-3 IP addresses are looked at, and the decapsulated packet is routed between subnets.
Layer-3 Switches
The other scenario I haven't covered is the Layer 3 switch. These switches operate the same way as described above. They are switches, but they can have specific ports, physical or virtual, designated as routed ports. These routed ports operate as gateways between subnets, (often VLANs within a switch) while the switched ports continue to only operate at Layer-2.
Part of the distinction between layer-2 and layer-3 devices is reflected in the hardware and memory of the devices. As Cisco explains, layer-2 only devices have a CAM (content addressable memory) table containing the MAC-Address-table. Layer-3 devices additionaly have a TCAM table, which handles mapping between routing, layer-2 and layer-3 addresses. Because of the physical hardware differences, you will see price differences in hardware that is a pure layer-2 switch, a layer-3 switch, and a router.
Best Answer
Correct. Apply the mask and you can see which part belongs to the network, and the reminder is the host address.
e.g. mask 255.255.255.0 and 1.2.3.4 would yield
It often does, but part of that is habitual. I could use a netmask of 255.255.255.128 (Binary 11111111.11111111.11111111.1000000) to get smaller subnets. And 10.20.30.128 would be a network address, with 10.20.30.129 the first possible IP on it, 10.20.30.130 the second possible etc etc.
And for 10.20.30.0 to 10.20.30.127 the broadcast IP would be 10.20.30.127.
Correct. They look at their routing table. If know have a specific route then they apply it. If not they forward it to their standard place. This is where the default gateway comes in.
No. Network address are usually not assigned. They could be but then things would break down in most cases because the network address used to be used as the broadcast address before this got changed to the highest address in the network. Different implementations of IP stack would make using the network adress fun.
You might get away with it in very specific circumstances. E.g. on a specifc OS and with specific routers. But once you start to mix things, upgrade drivers or send things outside of your control (e.g. onto the Internet) you are going to have job security while debugging...
That is merely a normal IP. 192.168.0.0/16 is usually used as a /24, with usuable IPs from 192.168.0.1 though 192.168.255.254 (65536-2 IPs). With the first IP usually reserved and the last IP reserved (as broadcast address).
Most SoHo setups take the RFC1918 range of 192.168.0.0/16, split in into 256 smaller subnets (/24) and assign the router the first or the last usable IP.
This makes a lot of sense when you design a network. One of the first things you need to plan the a way off the local network, and giving that the first IP got to be custom. But that is pure custom. Giving the router and IP of 192.168.1.42 is equally valid.
True.
Not sure about the versus. It contains network addresses. Briefly it does this:
Only for a /16. There are 30 possible network masks (excluding 0.0.0.0 and host-to-host specific ones) and only 3 of those end on the common round .0's