Question

A common protocol stack used by many applications is to use TCP at the transport layer and IP at the network layer. The ICMP protocol is also used to report on problems encountered by IP.

1. The IP protocol provides a service referred to as “best effort”. Describe the services of the IP protocol, and potentially ICMP protocol, with respect to flow control (ensuring the sender does not overwhelm the receiver).
2. Describe the service TCP provides with respect to flow control and explain the mechanisms it uses to overcome the deficiencies of IP.

Answer 1

Services of IP protocol:
The Internetwork Protocol (IP) provides a best effort network layer service for connecting computers to form a computer network. Each computer is identified by one or more globally unique IP addresses. The network layer PDUs are known as either "packets" or "datagram’s". Each packet carries the IP address of the sending computer and also the address of the intended recipient or recipients of the packet. Other management information is also carried.
The IP network service transmits datagram’s between intermediate nodes using IP routers. The routers themselves are simple since no information is stored concerning the datagram’s which are forwarded on a link. The most complex part of an IP router is concerned with determining the optimum link to use to reach each destination in a network. This process is known as "routing". Although this process is computationally intensive, it is only performed at periodic intervals.
An IP network normally uses a dynamic routing protocol to find alternate routes whenever a link becomes unavailable. This provides considerable robustness from the failure of either links or routers, but does not guarantee reliable delivery. Some applications are happy with this basic service and use a simple transport protocol known as the User Datagram Protocol (UDP) to access this best effort service.
Most Internet users need additional functions such as end-to-end error and sequence control to give a reliable service (equivalent to that provided by virtual circuits). This reliability is provided by the Transmission Control Protocol (TCP) which is used end-to-end across the Internet.
In a LAN environment, the protocol is normally carried by Ethernet, but for long distance links, other link protocols using fibre optic links are usually used. Other protocols associated with the IP network layer are the Internet Control Message Protocol (ICMP) and the Address Resolution Protocol (arp).
It provides services such as connection-oriented communication, reliability, flow control, and multiplexing. The details of implementation and semantics of the transport layer of the Internet protocol suite, which is the foundation of the Internet, and the OSI model of general networking are different.
Cisco IP Services comprises many basic and advanced building blocks that enable customers to: ... Deploy an IP network with basic end-to-end IP connectivity. Manage their IP addressing requirements from a central location. Control the IP addressing scheme used throughout their network.

Services of TCP :
Process-to-Process Communication –

TCP provides process to process communication, i.e, the transfer of data takes place between individual processes executing on end systems. This is done using port numbers or port addresses. Port numbers are 16 bit long that help identify which process is sending or receiving data on a host.
Stream oriented –

The data is sent and received as a stream of bytes(unlike UDP or IP that divides the bits into datagrams or packets). However, the network layer, that provides service for the TCP, sends packets of information not streams of bytes. Hence, TCP groups a number of bytes together into a segment and adds a header to each of these segments and then delivers these segments to the network layer. At the network layer, each of these segments are encapsulated in an IP packet for transmission. The TCP header has information that is required for control purpose which will be discussed along with the segment structure.
Full duplex service –

The communication can take place in both directions at the same time.
Connection oriented service –

Unlike UDP, TCP provides connection oriented service. It defines 3 different phases:
Connection establishment
Data transfer
Connection termination
(IMP: This is a virtual connection, not a physical connection, means during the transmission the resources will not be reserved and the segments will not follow the same path to reach the destination but it is a connection orientation in the sense that segments will arrive in order by the help of sequence number.)
Reliability –

TCP is reliable as it uses checksum for error detection, attempts to recover lost or corrupted packets by re-transmission, acknowledgement policy and timers. It uses features like byte number and sequence number and acknowledgement number so as to ensure reliability. Also, it uses congestion control mechanisms.
Multiplexing –

TCP does multiplexing and de-multiplexing at the sender and receiver ends respectively as a number of logical connections can be established between port numbers over a physical connection.
Mechanism it uses to overcome the deficiencies of IP:
TCP/IP was originally designed, and is usually implemented, for wide-area networks. While TCP/IP is usable on a local-area network, it is not optimized for this domain. For example, TCP uses an in-packet checksum for end-to-end reliability, despite the presence of per-packet CRC's in most modern network hardware. But computing this checksum is expensive, creating a bottleneck in packet processing. IP uses header fields such as `Time-To-Live' which are only relevant in a wide-area environment. IP also supports internetwork routing and in-flight packet fragmentation and reassembly, features which are not useful in a local-area environment. The TCP/IP model assumes communication between autonomous machines that cooperate only minimally. However, machines on a local-area network frequently share a common administrative service, a common file system, and a common user base. It should be possible to extend this commonality and cooperation into the network communication software.
Current TCP/IP implementations use a complicated memory management mechanism. This system exists for a number of reasons. First, a multi-layered protocol stack means packet headers are added (or removed) as the packet moves downward (or upward) through the stack. This should be done easily and efficiently, without excessive copying. Second, buffer memory inside the operating system kernel is a scarce resource; it must be managed in a space-efficient fashion. This is especially true for older systems with limited physical memory.
To meet these two requirements, mechanisms such as the Berkeley Unix mbuf have been used. An mbuf can directly hold a small amount of data, and mbufs can be chained to manage larger data sets. Chaining makes adding and removing packet headers easy. The mbuf abstraction is not cheap, however: 15% of the processing time for small TCP packets is consumed by mbuf management. Additionally, to take advantage of the mbuf abstraction, user data must be copied into and out of mbufs, which consumes even more time in the data transfer critical path. This copying means that nearly one-quarter of the small-packet processing time in a commercial TCP/IP stack is spent on memory management issues. Reducing the overhead of memory management is therefore critical to improving communications performance.