Quality of service at the computer networks based on internet
А framework for the emerging computer networks quality of Internet service is presented. Two important components of the framework are considered: integrated service and differentiated service. They are described and problems related to their implementation are discussed. Two broadly classified algo...
Gespeichert in:
| Veröffentlicht in: | Системні дослідження та інформаційні технології |
|---|---|
| Datum: | 2006 |
| Hauptverfasser: | , , |
| Format: | Artikel |
| Sprache: | English |
| Veröffentlicht: |
Навчально-науковий комплекс "Інститут прикладного системного аналізу" НТУУ "КПІ" МОН та НАН України
2006
|
| Schlagworte: | |
| Online Zugang: | https://nasplib.isofts.kiev.ua/handle/123456789/42200 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Zitieren: | Quality of service at the computer networks based on internet / N.I. Listopad, A.G. Kopachev, A.A. Matruk // Систем. дослідж. та інформ. технології. — 2006. — № 4. — С. 71–76. — Бібліогр.: 8 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
nasplib_isofts_kiev_ua-123456789-42200 |
|---|---|
| record_format |
dspace |
| spelling |
Listopad, N.I. Kopachev, A.G. Matruk, A.A. 2013-03-12T20:05:58Z 2013-03-12T20:05:58Z 2006 Quality of service at the computer networks based on internet / N.I. Listopad, A.G. Kopachev, A.A. Matruk // Систем. дослідж. та інформ. технології. — 2006. — № 4. — С. 71–76. — Бібліогр.: 8 назв. — англ. 1681–6048 https://nasplib.isofts.kiev.ua/handle/123456789/42200 621.321.1.519.872 А framework for the emerging computer networks quality of Internet service is presented. Two important components of the framework are considered: integrated service and differentiated service. They are described and problems related to their implementation are discussed. Two broadly classified algorithms are investigated: scheduling and queue management algorithms. The model of the average time delay for the type packet is presented. Запропоновано структуру забезпечення якості надання послуг Internet. Розглянуто два важливі компоненти цієї структури: інтегроване та диференційоване обслуговування, їх реалізацію та пов’язані з нею проблеми. Досліджено два алгоритма: планування і управління чергою. Запропоновано модель даних, яка базується на теорії масового обслуговування. Предложена структура обеспечения качества предоставления услуг Internet. Рассмотрены два важных компонента этой структуры: интегрированное и дифференцированное обслуживание, их реализация и связанные с ней проблемы. Исследованы два алгоритма: планирования и управления очередью. Предложена модель данных, базирующаяся на теории массового обслуживания. en Навчально-науковий комплекс "Інститут прикладного системного аналізу" НТУУ "КПІ" МОН та НАН України Системні дослідження та інформаційні технології Проблемно і функціонально орієнтовані комп’ютерні системи та мережі Quality of service at the computer networks based on internet Якість обслуговування у комп’ютерних мережах на базі Internet Качество обслуживания в компьютерных сетях на базе Internet Article published earlier |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| title |
Quality of service at the computer networks based on internet |
| spellingShingle |
Quality of service at the computer networks based on internet Listopad, N.I. Kopachev, A.G. Matruk, A.A. Проблемно і функціонально орієнтовані комп’ютерні системи та мережі |
| title_short |
Quality of service at the computer networks based on internet |
| title_full |
Quality of service at the computer networks based on internet |
| title_fullStr |
Quality of service at the computer networks based on internet |
| title_full_unstemmed |
Quality of service at the computer networks based on internet |
| title_sort |
quality of service at the computer networks based on internet |
| author |
Listopad, N.I. Kopachev, A.G. Matruk, A.A. |
| author_facet |
Listopad, N.I. Kopachev, A.G. Matruk, A.A. |
| topic |
Проблемно і функціонально орієнтовані комп’ютерні системи та мережі |
| topic_facet |
Проблемно і функціонально орієнтовані комп’ютерні системи та мережі |
| publishDate |
2006 |
| language |
English |
| container_title |
Системні дослідження та інформаційні технології |
| publisher |
Навчально-науковий комплекс "Інститут прикладного системного аналізу" НТУУ "КПІ" МОН та НАН України |
| format |
Article |
| title_alt |
Якість обслуговування у комп’ютерних мережах на базі Internet Качество обслуживания в компьютерных сетях на базе Internet |
| description |
А framework for the emerging computer networks quality of Internet service is presented. Two important components of the framework are considered: integrated service and differentiated service. They are described and problems related to their implementation are discussed. Two broadly classified algorithms are investigated: scheduling and queue management algorithms. The model of the average time delay for the type packet is presented.
Запропоновано структуру забезпечення якості надання послуг Internet. Розглянуто два важливі компоненти цієї структури: інтегроване та диференційоване обслуговування, їх реалізацію та пов’язані з нею проблеми. Досліджено два алгоритма: планування і управління чергою. Запропоновано модель даних, яка базується на теорії масового обслуговування.
Предложена структура обеспечения качества предоставления услуг Internet. Рассмотрены два важных компонента этой структуры: интегрированное и дифференцированное обслуживание, их реализация и связанные с ней проблемы. Исследованы два алгоритма: планирования и управления очередью. Предложена модель данных, базирующаяся на теории массового обслуживания.
|
| issn |
1681–6048 |
| url |
https://nasplib.isofts.kiev.ua/handle/123456789/42200 |
| citation_txt |
Quality of service at the computer networks based on internet / N.I. Listopad, A.G. Kopachev, A.A. Matruk // Систем. дослідж. та інформ. технології. — 2006. — № 4. — С. 71–76. — Бібліогр.: 8 назв. — англ. |
| work_keys_str_mv |
AT listopadni qualityofserviceatthecomputernetworksbasedoninternet AT kopachevag qualityofserviceatthecomputernetworksbasedoninternet AT matrukaa qualityofserviceatthecomputernetworksbasedoninternet AT listopadni âkístʹobslugovuvannâukompûternihmerežahnabazíinternet AT kopachevag âkístʹobslugovuvannâukompûternihmerežahnabazíinternet AT matrukaa âkístʹobslugovuvannâukompûternihmerežahnabazíinternet AT listopadni kačestvoobsluživaniâvkompʹûternyhsetâhnabazeinternet AT kopachevag kačestvoobsluživaniâvkompʹûternyhsetâhnabazeinternet AT matrukaa kačestvoobsluživaniâvkompʹûternyhsetâhnabazeinternet |
| first_indexed |
2025-11-24T03:23:44Z |
| last_indexed |
2025-11-24T03:23:44Z |
| _version_ |
1850840685665583104 |
| fulltext |
© N.I. Listopad, A.G. Kopachev, A.A. Matruk, 2006
Системні дослідження та інформаційні технології, 2006, № 4 71
UDC 621.321.1.519.872
QUALITY OF SERVICE AT THE COMPUTER NETWORKS
BASED ON INTERNET
N.I. LISTOPAD, A.G. KOPACHEV, A.A. MATRUK
А framework for the emerging computer networks quality of Internet service is pre-
sented. Two important components of the framework are considered: integrated ser-
vice and differentiated service. They are described and problems related to their im-
plementation are discussed. Two broadly classified algorithms are investigated:
scheduling and queue management algorithms. The model of the average time delay
τι for the type i packet is presented.
Today’s Internet only provides best effort services when traffic is processed as
quickly as possible, but there is no any guarantee as to timeliness or actual deliv-
ery, with the rapid transformation of the Internet into a commercial infrastructure,
demands for service quality have developed [1–3].
It is becoming apparent that several service classes are much demanded. One
of that service class will provide predictable Internet services for companies that
do business on the web. These companies are going to pay a certain price to make
their services reliable and to give their users high speed access for their web sites.
Another service class will provide low delay and low jitter services to applica-
tions such as Internet telephony and video conferencing. Companies are going to
pay a premium price to run a high quality videoconference to save travel time and
other costs. Finally, the best effort service will remain for those customers who
need only connectivity.
Which mechanism should be chosen to provide QOS is a very topical issue.
There is the opinion which asserts that fibers and wavelength division multiplex-
ing (WDM) will make bandwidth so abundant and cheap that QOS will be auto-
matically delivered. The other opinion asserts that no matter how much band-
width the network can provide, new application will be invented to consume
them.
Therefore mechanisms will still be needed to provide QOS. Here you should
simply note that, even if bandwidth will eventually is not going to happen soon,
for now, some simple mechanisms are definitely needed in order to provide QOS
on the Internet [2].
The Internet engineering task force (IETF) has proposed many service mod-
els and mechanisms to meet the demand for QOS. Among them are the integrated
services/RSVP model, the differentiated services (DS) model, traffic engineering
and constraint based routing [1,3].
The integrated service model is characterized by resource reservation for real
time applications, before data are transmitted. The applications must first set up
paths and reserve resources. So, RSVP is a signaling protocol for setting up paths
and reserving resources.
The integrated service model proposes two service classes in addition to best
effort service [2, 3]. They are:
N.I. Listopad, A.G. Kopachev, A.A. Matruk
ISSN 1681–6048 System Research & Information Technologies, 2006, № 4 72
1) guaranteed service: for applications requiring fixed delay bound;
2) controlled load service: for applications requiring reliable and enhanced
best effort service.
The paradigm of this model is that «there is an inescapable requirement for
routers to be able to reserve resources in order to provide special QOS for specific
user packet streams, or flows. This in turn requires flow-specific state in the
routers».
As mentioned above RSVP was developed as a signaling protocol for appli-
cations to reserve resources. The signaling process is illustrated in fig. 1: the
sender sends a PATH message to the receiver specifying the characteristics of the
traffic. Every intermediate router along the path forwards the PATH message to
the next hop determined by the routing protocol. Upon receiving a PATH mes-
sage, the receiver responds with a RESV message to request resources for the
flow. Every intermediate router along the path can reject or accept the request of
the RESV message. If the request is rejected, the router will send an error mes-
sage to the receiver, and the signaling process will terminate. If the request is ac-
cepted, link bandwidth and buffer space are allocated for the flow and the related
flow state information will be installed in the router.
Integrated services are implemented by four components: the signaling pro-
tocol (e.g. RSVP), the admission control routine, the classifier and the packet
scheduler. Applications requiring guaranteed service or control- load service must
set up the paths and reserve resources before transmitting their data. The admis-
sion control routines, will decide whether a request for resources can be granted.
When a router receives a packet, the classifier will perform a multi-field (MF)
classification and put the packet in a specific queue based on the classification
result. The packet scheduler then schedules the packet according to QOS re-
quirements.
The integrated services/RSVP architecture represents a fundamental change
to the current internet architecture, which is based on the concept that all flow-
related state information should be in the end systems.
The problems with the integrated services architecture are the following [3].
1. The amount of state information increases proportionally with the number
of flows. This places a huge storage and processing overhead on the routers,
therefore, this architecture does not scale well in the Internet core.
2. The requirements on the router are high. All routers must implement
RSVP, admission control, MF classification and packet scheduling.
3. Ubiquitous deployment is required for guaranteed service. Incremental
deployment of controlled-load is possible by deploying controlled-load service
Fig. 1
Sender Receiver Router Router
(6) RESV (4) RESV
(1) PATH (3) PATH
(2) PATH
(5) RESV
Quality of service at the computer networks based on Internet
Системні дослідження та інформаційні технології, 2006, № 4 73
and RSVP functionality at the bottleneck nodes of a domain and tunneling the
RSVP messages over other part of the domain.
Due to the difficulty in implementing and deploying integrated services and
RSVP, differentiated services (DS) is introduced. IPv4 header contains a TOS
(type of service) byte. Applications can set three bits in the TOS byte to indicate
the need for low delay or low loss rate service. However, choices are limited. Dif-
ferentiated service defines the layout of the TOS byte (termed DS field) and a
base set of packet forwarding treatments (termed per-hop behaviors, or PHBs). By
marking the DS fields of packets differently, and handling packets based on their
DS fields, several differentiated service classes can be created. Therefore, differ-
entiated services form a relative-priority scheme.
Customers can mark DS fields of individual packets to indicate the desired
service or have them marked by the leaf router based on MF classifications at the
ingress of the ISP (Internet service provider) networks; packets are classified, po-
liced and possibly shaped. The classification, policing and shaping rules used at
the ingress routers are driven from the SLAs (service level agreements). The
amount of buffering space needed for these operations is also driven from the
SLAs. When the packet inters one domain from another domain, its DS field may
be re-marked as determined by the SLA between the two domains.
Differentiated service is different from integrated service. First, there are
only a limited number of service classes indicated by the DS field. Since service
is allocated in granularity of a class, the amount of state information is propor-
tional to the number of classes rather than the number of flows. Differentiated
service is therefore more scalable. Second, sophisticated classification, marking,
policing and shaping operations are only needed at boundary of the networks. ISP
core routers need only to implement behavior aggregate (BA) classification.
Therefore, it is easier to implement and deploy differentiated services.
There is another reason why the second feature is desirable for ISPs. ISP
networks usually consists of boundary routers connected to customers and core
routers/switches interconnecting the boundary routers. Core routers must forward
packets very fast and therefore must be simple. Boundary routers need not for-
ward packets very fast because customer links are relatively slow. Therefore, they
can spend more time on sophisticated classification, policing and shaping.
Boundary routers at the network access points (NAPs) are exceptions. They must
forward packets very fast and so sophisticated classification, policing and shap-
ing. Therefore, they must be well equipped.
In the differentiated services model incremental deployment is possible for
assured service. DS-incapable routers simply ignore the DS fields of the packets
and give the assured service packets best effort service. Since assured service
packets are less likely to be dropped by DS-capable routers, the overall perform-
ance of assured service traffic will be better than the best effort traffic.
In this paper a service architecture for differentiated service is presented.
The architecture provides assured service, premium service in addition to best
effort service.
Assured service is intended for customers that need reliable services for their
service providers, even in time of network congestion. Customers will have SLAs
with their ISPs. The SLAs will specify the amount of bandwidth allocated for the
customers. Customers are responsible for deciding how their applications share
N.I. Listopad, A.G. Kopachev, A.A. Matruk
ISSN 1681–6048 System Research & Information Technologies, 2006, № 4 74
that amount of bandwidth. SLAs for assured service are usually static, meaning
that the customers can start data transmission whenever they want without signal-
ing their ISPs.
At present there are two directions of investigating broadly classified under
the monikers «scheduling algorithms» and «queue management algorithms» [2].
The generic scheduling algorithm, exemplified by the well-known Fair Queuing
(FQ) algorithm, requires the buffer at each output of a router to be partitioned into
separate queues each of which will buffer the packets of one of the flows. Packets
from the flow buffers are placed on the outgoing line by a scheduler using an ap-
proximate bit-by-bit, round-robin discipline. Because of per flow queuing, pack-
ets belonging to different flows are essentially isolated from each other and one
flow cannot degrade the quality of another. However, it is well-known that this
approach requires complicated per flow state information, making it too expen-
sive to be widely deployed.
To reduce the cost of maintaining flow state information, I.Stoica has re-
cently proposed a scheduling algorithm called Core Stateless Fair Queuing
(CSFQ) [4]. In this method routers are divided into two categories: edge routers
and core routers. An edge router keeps per flow state information and estimates
each flow’s arrival rate. These estimates are inserted into the packet headers and
passed on to the core routers. A core router simply maintains a stateless FIFO
queue and, during periods of congestion, drops a packet randomly based on the
rate estimates. This scheme reduces the core router’s design complexity. How-
ever, the edge router’s design is still complicated. Also, because of the rate infor-
mation in the header, the core routers have to extract packet information differ-
ently from traditional routers.
Another notable scheme which aims to approximate FQ at a smaller imple-
mentation cost is Stochastic Fair Queuing (SFQ) proposed by McKenny [5]. SFQ
classifies packets into a smaller number of queues than FQ using a hash function.
Although this reduces FQ’s design complexity, SFQ still requires around 1000 to
2000 queues in a typical router to approach FQ’s performance.
Other directions of algorithms, queue management algorithms have had a
simple design from the outset. Given their simplicity, the hope is to approximate
fairness. This class of algorithms is exemplified by Random Early Detection
(RED) proposed by S. Floyd and V. Jacobsen [5, 6]. A router implementing RED
maintains a single FIFO to be shared by all the flows, and drops an arriving
packet at random during periods of congestion. The drop probability increases
with the level of congestion. Since RED acts in anticipation of congestion, it does
not suffer from the «lock out» and «full queue» problems inherent in the widely
deployed Drop Tail mechanism. By keeping the average queue-size small, RED
reduces the delays experienced by most flows. However, like Drop Tail, RED is
unable to penalize unresponsive flows, which are based on UDP datagram proto-
col. This is because the percentage of packets dropped from each flow over a pe-
riod of time is almost the same. Consequently, misbehaving traffic can take up a
large percentage of the link bandwidth and starve out TCP friendly flows and by
this it does harm to the end users.
Premium service provides low-delay and low jitter service for customers that
generate fixed peak bit- rate traffic. Each customer will have a SLA with its ISP.
The SLA specifies a desired peak bit- rate for a specific flow or an aggregation of
Quality of service at the computer networks based on Internet
Системні дослідження та інформаційні технології, 2006, № 4 75
flows. The customer is responsible for not exceeding the peak rate. Otherwise,
excess traffic will be dropped. The ISP guarantees that the contracted bandwidth
will be available when traffic is sent. Premium service is suitable for Internet te-
lephony, video conferencing, or for creating virtual lease lines for private net-
works (VPNs).
Because premium service is more expensive than assured service, it is desir-
able for ISPs to support both static SLAs and dynamic SLAs. Dynamic SLAs allow
customers to request for premium service on demand without subscribing to it.
Admission control is needed for dynamic SLAs.
Premium service can be implemented as follows. At the customer side, some
entity decides which application flow can use premium service. The leaf routers
directly connected to the senders will do MF classifications and shape the traffic.
Conceptually, we can consider that there is a P-bit in the DS field. If the P-bit of a
packet is set, this packet belongs to the premium class. Otherwise, the packet be-
longs to the assured service class or best effort class. After the shaping, the P-bits
of all packets are set for the flow that is allowed to use premium service. The exit
routers of the customer domain may need to reshape the traffic to make sure that
the traffic does not exceed the peak rate specified by the SLA. At the provider
side, the ingress routers will police the traffic. Excess traffic is dropped and all
packets with the P-bit set enter a premium queue (PQ). Packet in the PQ will be
sent before packets in the AQ.
By limiting the total amount of bandwidth requested by premium traffic, the
network administrators can guarantee that premium traffic will not starve the as-
sured and best effort traffic.
To improve the required QoS we are supposed to solve the following tasks:
• Determining network congestion.
• Providing dedicated bandwidth for media flows.
• Providing individual network characteristics for given media flow.
At present there are two directions of investigating broadly classified under
the monikers «scheduling algorithms» and «queue management algorithms» All
of the router algorithms (scheduling and queue management) developed so far
have been either able to provide fairness or simplicity to implement, but not both
features simultaneously. We will take a step in the direction of bridging fairness
and simplicity. Specifically, I have an idea to exhibit an active queue management
algorithm, that could be simple to implement (since it doesn’t requires state in-
formation) and could differentially penalize misbehaving.
The tasks to be proposed.
• Developing a simple, stateless algorithms that could achieve flow isola-
tion and/or approximate fair bandwidth allocation and looking for a solution to
the above problem in the context of the IP based networks. Also we need to find
scheme that could differentially penalize «unresponsive» flows.
• Testing our research on high speed network based on the developed
methods and algorithms. For evaluating the effect of the proposed methods as the
base model for communication link we are going to use the mass servicing system
with the heterogenous packets flows on n types, which get into the channel with
the intensities nλλ ,...,1 . Let V — the total link bandwidth, iL — the average
N.I. Listopad, A.G. Kopachev, A.A. Matruk
ISSN 1681–6048 System Research & Information Technologies, 2006, № 4 76
length of the type i packet. When the packets are of the same length and flows are
simple, the average time delay for the type i packet will be calculated as follow-
ing:
V
L
LVLV
L
i
i
j
i
j
iiii
n
j
jj
i +
−−
=
∑ ∑
∑
−
= =
=
1
1 1
1
2
))(( λλ
λ
τ , ni ,1= .
All the mathematical calculations are supposed to be compared with the
practical results of the algorithms.
Among the existing network technologies Asynchronous Transfer Mode
(ATM) meets the QOS requirements. Nevertheless ATM technology is rather ex-
pensive and complicated, especially, for using in R&D networks.
REFERENCES
1. Mieghem P. Van, De Neve H. and Kuipers F.A. Hop-by-hop quality of service rout-
ing // Computer Networks. — October 2001. — 37, № 3–4. — P. 423 – 497.
2. Копачев А.Г. Обзор методов и технологий по обеспечению качества предостав-
ляемых услуг в компьютерных сетях передачи данных // Информатизация
образования. — 2004. — № 4. –– С. 59–70.
3. Матрук А.А. Качество обслуживания в компьютерных сетях // Информатиза-
ция образования. — 2005. — № 3. — С. 81–83.
4. Stoica I., Zhang H. LIRA: A model for service differentiation in the Internet // In
proc. of NOSSDAV’98. — London: UK. — 1998.
5. Floyd S., Jacobson V. Link sharing and resource management models for
packet networks // IEEE/ACM transactions on networking. — 1995. — 3,
№ 4. — P. 365–386.
6. Floyd S., Jacobson V. Random early detection gateways for congestion avoidance //
IEEE / ACM transaction on networking. — 1993. — 1, № 4. — P. 397–413.
Received 15.03.2006
|