Professional Documents
Culture Documents
Traffic Analysis in The TRAMMS Project
Traffic Analysis in The TRAMMS Project
customers. The FTTH customers represent many social the other networks. The main conclusion is that the shape of
and ethnic groups, while the DSL customers constitute a the daily traffic patterns depends on the subscriber type of
more homogeneous group of Swedish middle class living the network such as residential, enterprise or academic. There
in single family houses. is also a common daily traffic pattern for the networks that
• Swedish municipal network 2: a FTTH network with 350 have mainly residential users which is the same for both the
IPTV users. Measurements from this network was used Swedish and the Spanish residential networks.
for studying IPTV user behavior.
• Spanish commercial network: a network containing both 1
RedIRIS
fixed and wireless access networks. The wireline part 0.9 CMTS
FTTH
0.8 DSL
consists of a fibre network to the cabinet (FTTC) and the 0.7
Average
Figure 1. Normalized daily downlink traffic pattern for the different networks
A. Measurement tools and normalized average daily downlink traffic pattern.
The measurements in the residential networks have been
performed using PacketLogic (PL) [2], a commercial real-time
hardware/software solution used mainly for traffic surveil- B. Comparison of application usage in different networks and
lance, traffic shaping or as a firewall. Traffic is identified based technologies
on packet content and flow behavior using deep packet in- A comparison of application usage in the different net-
spection and deep flow inspection. PL uses the self-developed works was performed in order to see differences between
Datastream Recognition Definition Language (DRDL) [3] to user groups and countries. This comparison uses data from
identify different application protocols. The PL can identify Swedish municipal network 1 and the Spanish commercial
more than 800 application protocols. networks. In order to avoid differences in the amount of
The measurements in the RedIRIS network have been unrecognized traffic between the networks, only the traffic
performed with Cisco NetFlow [4]. NetFlow is a proprietary from the following application groups has been considered:
network protocol developed by Cisco Systems to run on web browsing, P2P file sharing and multimedia streaming.
their routers and implemented by other vendors as well. This These application groups are seen to have the largest traffic
protocol is used to monitor the traffic that traverses a router and volumes.
to keep statistics of the performance by sampling some of the Concerning uplink traffic, P2P file sharing is responsible for
packets. Cisco defines a flow as a unidirectional sequence of more than 97% of the considered traffic in the fixed networks
packets sharing all the following 7 values, commonly referred regardless of the technology (FTTH, DSL and CMTS). In the
as 7-tuple: Source and Destination IP addresses, IP protocol, case of the downlink traffic, in the fixed networks (FTTH,
Source and Destination ports (when the IP protocol is TCP or DSL and CMTS) the P2P file sharing generates an important
UDP), Ingress interface and IP Type of Service. amount of traffic depending on the technology (from 66% to
88%). Approximately 60% of the rest of the traffic corresponds
IV. T RAFFIC ANALYSIS to web browsing and 40% to multimedia streaming. Regarding
the mobile network (GGSN) the amount of web browsing
The results in this paper are taken from the TRAMMS traffic is five times higher than the multimedia streaming.
public deliverable D3.2 [5], available at [1] on request. Compared to the fixed networks, the P2P file sharing traffic in
the mobile network is lower in uplink (77% of the considered
A. Daily profiles traffic) and much lower in downlink (27% of the considered
An overview of the normalized average daily profiles in the traffic). In downlink the mobile network traffic belongs mainly
networks is shown in Figure 1. to web browsing (61% of the traffic).
The figure shows that the Spanish fixed network and the Downlink (%)
Swedish network, despite using different access technologies Sw network 1 Sp network
(CMTS, DSL, FTTH), have similar daily traffic patterns in the App group FTTH DSL CMTS GGSN
Web browsing 7.06 20.6 12.1 60.5
downlink direction. The same similarity is seen for the uplink P2P file sharing 88.3 66.0 80.8 26.6
traffic as well. However, the RedIRIS academic network shows Multimedia streaming 4.7 13.4 7.0 13.0
a very different daily downlink traffic pattern. The amount of Table I
downlink traffic is less constant in the RedIRIS network than VOLUME SHARE OF APPLICATION GROUPS IN THE DIFFERENT NETWORKS
in the other networks (10% of the maximum traffic at 5 a.m.)
and from 1 p.m. to 9 p.m. it decreases while it increases in
3
Uplink (%)
Sw network 1 Sp network
App group FTTH DSL CMTS GGSN
Web browsing 0.4 2.6 1.9 20.7
P2P file sharing 99.4 96.7 97.0 76.5
Multimedia streaming 0.2 0.6 1.2 2.8
Table II
VOLUME SHARE OF APPLICATION GROUPS IN THE DIFFERENT NETWORKS
IDs and time stamps when the videos were viewed. The latter
is determined by the packet arrival time. The packet dump is
anonymized and processed to obtain a text file containing only
the content IDs and times. This text file is later loaded into a
database system for further analysis.
Figure 4 shows the number of viewed videos per hour
throughout the 16 day long measurement, which is an estima-
tion of the user activity (and the traffic intensity as well). The
user activity seems more intense on weekdays and lower on
the weekend (confirmed by Figure 5 as well). In Figure 5 the
(a) Packets
final day of the measurement is cut so that it contains samples
of exactly two weeks; this way no distortion is introduced in
the sampling. The same technique is applied to calculate the
daily distribution (see Figure 6 below) of the access times.
60
50
40
30
20
10
0
0h 12h 0h 12h 0h 12h 0h 12h 0h 12h 0h 12h 0h 12h 0h 12h
Time (from 2008. november 17., Mon, 16:31:26 to 2008. november 25., Tue, 18:11:02)
(b) Bytes
Figure 4. Number of viewed YouTube videos per hour (Swedish network
measurement (DSL) from 2008-11-17 to 2008-12-02)
Figure 3. Traffic in the incoming direction. Spain and the US are excluded
from the figure.
90
from Spain than from the United States although the number
Number of YouTube videos viewed (per 15 mins)
80
70
of bytes received from the United States were greater than
60 that received from Spain. Concerning the youtube content
50 analysis, it is found that the peak hour is consistent with the
40 general daily traffic pattern. Also, the video content popularity
30 is uneven with a few video clips that are very popular, while
20 the rest are seldom watched.
10
The TRAMMS project is ongoing and measurement data are
0
0 AM 6 AM 12 PM 18 PM 24 PM collected continuously. The analysis of these measurements
Time of the day
will, of course, go on and be deepened to gain more specific
knowledge concerning content and user behavior. The content
Figure 6. Daily distribution of viewed YouTube videos per hour (Swedish locality studies and youtube analysis previously described are
network measurement (DSL) from 2008-11-17 to 2008-12-02)
preliminary and will be further explored within the project. In
the future the aim is to include more of end user behavior and
6 user quality of experience.
5
VI. ACKNOWLEDGEMENTS
The work has partly been financed by the CELTIC project
Number of views
4
TRAMMS, with the Swedish Governmental Agency for Inno-
3 vation Systems (VINNOVA) supporting the Swedish contribu-
tion. Maria Kihl is financed in the VINNMER programme at
2
VINNOVA. UAM would like to acknowledge the support of
1
the DIOR project that gave us the RedIRIS measurements.
0 R EFERENCES
0 500 1000 1500 2000 2500
YouTube videos (ranked) [1] “Traffic measurements and models in multi-service networks.”
(a) linear-linear scale http://projects.celtic-initiative.org/tramms/.
[2] “Procera networks.” http://www.proceranetworks.com.
[3] “Packetlogic drdl signatures and properties 10340 (beta),” tech. rep., 2007.
6
[4] “Cisco netflow.” http://www.cisco.com/go/netflow.
[5] “Deliverable d3.2 - traffic models,” tech. rep., 2009.
5 [6] http://www.maxmind.com/app/geolitecountry.
[7] K.Papagiannaki, N. Taft, S. Bhattacharyya, P. Thiran, K. Salamatian, and
Number of views
0
0 1 2 3 4
10 10 10 10 10
YouTube videos (ranked)