The Internet technology was built on a foundation that did not require switched services
(so that connectivity was not established on per-call basis). Based on the inter-

net specification (IETF RFC 791), the internet protocol (IP) is defined as a best-effort
connectionless technology. Thus, IP does not establish a fixed (switched) path dedicated
for the duration of a session, but instead, it assembles packets and by store-and-forward
it delivers each one over one or more routes, taking advantage of the temporal
availability of bandwidth resources. To keep track of packets at the receiving end,
all data devices on the network are assigned an individual address, and the packet has
a source and a destination address, a packet sequence number, and other overhead information.
Thus, IP proved to be an economical delivery of data method that led to its
recent explosion, despite the lack of packet delivery guarantee or its timely delivery.

Initially, the store-and-forward method of data delivery by routers did not have
real-time constraints, and did not require network high performance and network
reliability metrics, similar to the public switched network. For example:

• Latency exceeded 500 ms; the acceptable round-trip delay in synchronous
  networks should be <300 ms. Round-trip delay is measured from phone-tophone
  and back, based on POTS service experience.
• Quality of service (QoS) was nonexistent and consisted of only what is
  known as best effort (meaning, just try and if it fails, it’s OK)
• Privacy was not part of the initial IP strategy

Although the IP packet length is variable and a maximum length is defined, the majority
of packets do not have the maximum length. As already discussed, the packet
length and packet frequency are key characteristics of traffic engineering and data
network performance. In fact, the IETF Internet Protocol Performance Metrics
(IETF IPPM) group and the cooperative association for Internet data analysis (CAIDA)
have defined IP performance metrics for evaluating a data network. A good
understanding of these metrics helps to develop strategies that can provide better
service. Among such metrics are:

• Symmetry or asymmetry characteristics of data flow (geographical, temporal,
  and protocol-related),
• Packet length distribution (predominant packet size on the IP network)
• Length of packet train or packet flow distribution (typical number of packets
  in a single transaction)
• Causes of packet delay
• Causes of traffic congestion on the network
• Protocols and their application on the IP network (TCP, http)

Currently, the two dominant protocols in use are the TCP and HTTP. The TCP/IP
protocol accounts for more than 90% of the traffic and the http protocol for more
than 75%. Figure 3.7 illustrates the TCP/IP frame.

The explosion of IP demanded more sophisticated and robust services and networks
with restoration and real-time delivery characteristics, so that voice and real-time
compressed video can be transmitted over the Internet network and from multiple
sources with multiple protocols. However, since all sources are not equal,
there is a trust rate assigned to each. Clearly, the most sophisticated source will
have the highest trust rate, such as a directly connected interface or a manually entered
static routing, and the least sophisticated, such as an internal border gateway
protocol, will have the lowest trust rate. Thus, the highest trusted-rate source starts
to look very similar to traditional synchronous communications networks. As a consequence,
encapsulated Internet is been deployed in “several forms” in the optical
network, such as, for example, “Internet over SONET,” “Internet over ATM over
SONET,” and so on, in an effort to offer QoS, service flexibility, granularity, scalability,
shared access, point-to-multipoint, customized bit rates, and easy service migration.

That is, the evolving IP network includes newer IP protocol versions, such as
IPv6, in an effort to support compressed Voice over IP (VoIP) as well as compressed
Video over IP; VoIP is also known as Telephony over IP and as IP Telephony.
Thus, the two telecommunications camps, “voice/TDM” and “data/packet,”
were racing each other to offer asynchronous services over the synchronous network
and synchronous services over the (packetized data) asynchronous network,

both claiming that their networking solution was better than the other, or that their
services afforded lower cost.