In many ways, this is reminiscent of the dramatic impact the web had on the internet when we published the 1st edition of the book in 1996. Another widespread application class of the Internet is the delivery of "streaming" audio and video.
Requirements
- Stakeholders
- Scalable Connectivity
- Cost-Effective Resource Sharing
- Support for Common Services
- Manageability
Also note that the cloud is one of the most important icons of computer networks. The final requirement is that each node must be able to say which of the other nodes in the network it wants to communicate with.
Architecture
- Layering and Protocols
- Encapsulation
- Multiplexing and Demultiplexing
- Internet Architecture
Consider what happens when one of the application programs sends a message to its peer by passing the message to RRP. A final attribute of the Internet architecture (or more specifically the IETF culture) is that for a new protocol to be officially incorporated into the architecture, there must be both a protocol specification and at least one (and preferably two) representative implementations of the specification.
Software
Application Programming Interface (Sockets)
Because most network protocols are in software (especially those high in the protocol stack) and almost all computer systems implement their network protocols as part of the operating system, when we talk about an interface. On the server computer, the application process runs apassiveopen—the server says it's ready to accept connections, but doesn't actually make the connection.
Example Application
This is a blocking operation that does not return until a remote participant has established a connection, and when it does, it returns a new socket corresponding to this newly established connection, and the address argument contains the remote participant's address. In practice, the client usually just specifies the remote participant's address and lets the system fill in the local information.
Performance
- Bandwidth and Latency
- Delay × Bandwidth Product
- High-Speed Networks
- Application Performance Needs
This is a function of the network bandwidth and the size of the packet in which the data is carried. On the other hand, if the sender does not fill the pipe—that is, does not send an entire CAT × bandwidth product's worth of data before it stops waiting for a signal—the sender will not fully utilize the network.
Perspective: Feature Velocity
If the packets arrive at the destination exactly 33 ms apart, we can conclude that the delay experienced by each packet in the network was exactly the same. The network's design is clearly evolving, and we've documented these changes with each new edition of the handbook over the years.
DIRECT LINKS
Problem: Connecting to a Network
Technology Landscape
Before diving into the challenges outlined in the problem statement at the beginning of this chapter, it is useful to first get a lay of the land, which includes a wide range of link technologies. At the other end of the spectrum, a typical user mostly encounters links as a way to connect a computer to the existing Internet.
Encoding
The problem with the Manchester coding scheme is that it doubles the speed at which the signal passes over the link, meaning that the receiver has half the time to detect each signal pulse. In the case of Manchester encoding, the bit rate is half the baud rate, so the encoding is considered only 50% efficient.
Framing
- Byte-Oriented Protocols (PPP)
- Bit-Oriented Protocols (HDLC)
- Clock-Based Framing (SONET)
This approach is often called character padding because additional characters are inserted into the data portion of the frame. One of the fields in the upper part of the frame indicates the beginning of the load.
Error Detection
- Internet Checksum Algorithm
- Cyclic Redundancy Check
One is to notify the sender that the message has been corrupted so that the sender can resend a copy of the message. Second, many errors will go undetected—any error that happens to corrupt the same bit positions in the first and second copies of the message.
Reliable Transmission
- Stop-and-Wait
- Sliding Window
- Concurrent Logical Channels
Timeline with four different scenarios for the stop-and-wait algorithm. a) The ACK is received before the timer expires; (b) the original list is lost; Then, should frame 6 arrive - perhaps it's too late because it was lost the first time and had to be retransmitted, or maybe it was just delayed - the receiver acknowledges frame 8, raises LFR to 8 and sets LAF to 12.1. If frame 6 is indeed lost, the sender has timed out, causing it to retransmit frame 6. uh.
Multi-Access Networks
- Physical Properties
- Access Protocol
- Longevity of Ethernet
Technically, the address belongs to the adapter, not the host; it is usually written in ROM. The adapter gives up after a certain number of attempts and reports a transfer error to the host. First, Ethernet is very easy to manage and maintain: there are no routing or configuration tables to update, and it's easy to add a new host to the network.
Wireless Networks
- Basic Issues
- Bluetooth (802.15.1)
This is despite the fact that radio waves emitted by one client node can be received by other client nodes - the shared base station model does not allow direct communication between client nodes. This effectively tells nodes within range of the receiver that they should not send anything for a while - the amount of target transmission time is included in the RTS and CTS packets. Most of the 802.11 frame format, which is shown in Figure 2.30, is exactly what we would expect.
Access Networks
- Passive Optical Network
- Cellular Network
In the case of a mobile network, base stations are often called broadband base units (BBUs), the mobile devices that connect to them are usually called user equipment (UEs), and the set of BBUs is anchored in an evolved packet core (EPC) hosted at a central office. In essence, 5G defines a family of waveforms—unlike 4G, which defined only one waveform—each optimized for a different band in the radio spectrum.2 Bands with carrier frequencies below 1 GHz are designed to deliver mobile broadband and large-scale IoT services with a primary focus on range. These different waveforms affect the scheduling and spacing of the subcarriers (ie the "size" of the source elements just described).
Perspective: Race to the Edge
In this scenario, the access network remains a dumb bit-pipe that allows cloud providers to excel at what they do best: run scalable cloud services on commodity hardware. On the other hand, network operators believe that by building next-generation access networks using cloud technology, they will be able to co-locate edge applications in the access network. The idea is to make the access-edge cloud available to everyone, and not exclusively the domain of established cloud providers or network operators.
THREE
INTERNETWORKING
Problem: Not All Networks are Directly Connected
Switching Basics
- Datagrams
- Virtual Circuit Switching
- Source Routing
In the connection configuration phase, it is necessary to establish a "connection state" in each of the switches between the source and destination hosts. For each packet it wants to send to host B, A puts the VCI value of 5 in the packet's header and sends it to switch 1. One would be to assign a number to each output of each switch and put that number in the packet's header.
Switched Ethernet
- Learning Bridges
- Implementation
- Spanning Tree Algorithm
- Broadcast and Multicast
- Virtual LANs (VLANs)
When a switch receives a configuration message indicating that it is not the root—that is, a message from a switch with a smaller ID—the switch stops generating configuration messages on its own and instead forwards configuration messages from other switches, after first adding 1 to the distance field. Similarly, when a switch receives a configuration message indicating that it is not the designated switch for that port—that is, a message from a switch that is closer to the root or equidistant from the root but with a smaller ID—the switch stops sending configuration messages over that port. Note that hosts A and B cannot communicate via the shortest path (via S5) because frames must "flow up the tree and back down", but that's the price you pay to avoid loops.
Internet (IP)
- What Is an Internetwork?
- Service Model
- Global Addresses
- Datagram Forwarding in IP
- Subnetting and Classless Addressing
- Address Translation (ARP)
- Host Configuration (DHCP)
- Error Reporting (ICMP)
- Virtual Networks and Tunnels
A reasonable choice is the MTU of the network to which the host is directly connected. The network portion of an IP address uniquely identifies a single physical network that is part of the larger Internet. In the routing table of the router at the entrance of the tunnel, this virtual link looks a lot like a normal link.
Routing
- Network as a Graph
- Distance-Vector (RIP)
- Link State (OSPF)
- Metrics
The initial assumption for distance vector routing is that each node knows the cost of connecting to each of its directly connected neighbors. Each node is supposed to be able to detect the link state with its neighbors (up or down) and the cost of each link. A list of that node's directly connected neighbors, with the cost of connecting to each.
Implementation
- Software Switch
- Hardware Switch
- Software Defined Networks
The non-trivial algorithms discussed in this chapter—the spanning tree algorithm used by learning bridges, the distance vector algorithm used by RIP, and the link state algorithm used by OSPF—are not directly part of the per-packet forwarding decision. Just as importantly, the full architectural specification for switches that take advantage of these new processors is now available online—the hardware equivalent of open source software. Exactly how one "programs" the NPU depends on the chip vendor, of which there are currently several.
Perspective: Virtual Networks All the Way Down
Similarly, server virtualization represents an abstraction of a virtual machine (VM) that has all the properties of a physical machine. If you want to change the purpose of the World Turtle mythology: it's virtual networks all the way. Addressing this challenge will be at the heart of networking for the next decade, and while some of this work will undoubtedly be done in proprietary settings, there are open source network virtualization platforms (eg, the Linux Foundation's Tungsten Fabricproject) leading the way.
FOUR
ADVANCED INTERNETWORKING
Problem: Scaling to Billions
Global Internet
- Routing Areas
- Interdomain Routing (BGP)
All routers in the area send each other link state advertisements and thus develop a complete and stable map of the area. Since another name for autonomous systems on the Internet is domain routing, we refer to the two parts of the routing problem as interdomain routing and intradomain routing. Provider-Client - Providers are in the business of connecting their clients to the rest of the Internet.
IP Version 6
- Historical Perspective
- Addresses and Routing
At the same time, each router in the AS keeps track of how to get to each border router using a conventional intradomain protocol with no information injected. In this way, each router in the AS can compile a complete routing table for each prefix that is reachable through a border router of the AS. This means a new version of the Internet protocol and, as a result, new software for every host and router on the Internet.