Fleksibilni Internet servisi na bazi kontrole kašnjenja i

[12]-[23] and loss rate differentiation [15], [24]-[27] among classes emerged recently.
However, combined service differentiation across multiple QoS metrics is still an open
issue because these mechanisms are not applicable to flows whose performance depends
both on delay and loss rate, such as TCP flows. In case of TCP flows, delay and loss rate
differentiations cannot be performed independently because their combined effect on
TCP throughput might be unpredictable. Introduction of a low-delay service for delay-
sensitive packets might increase the queuing delay and round-trip time for TCP packets
that are used for bulk data transfer. Since the throughput of TCP flows depends on the
round-trip time, delay differentiation might have negative effect on TCP applications.
Since TCP traffic represents up to 95 % of the total traffic in today’s Internet [28],
approaches that do not consider the interaction between service differentiation
mechanism and TCP are not likely to be successful in practice.
In this thesis, we propose a new non-elevated service differentiation architecture
that provides delay and throughput differentiation between delay-sensitive and
throughput-sensitive traffic classes. An Internet application may choose to mark its
packets either as throughput-sensitive or delay-sensitive by setting a field in packets’ IP
headers. Both classes benefit from the proposed architecture because delay-sensitive
(real-time) applications receive low-delay service, while throughput-sensitive (TCP)
applications are guarantied at least the same throughput as they would receive in a best-
effort network. In addition to the new algorithm named Packet Loss Controller (PLC)
[30], the proposed architecture employs two existing router algorithms: a scheduling
algorithm for proportional delay differentiation called Backlog-Proportional Rate (BPR)
[12] and an active queue management (AQM) called Random Early Detection (RED)
3