Who Runs the Internet?
Classifying Autonomous Systems into Industries
Annika Baumann and Benjamin Fabian
Institute of Information Systems, Humboldt University Berlin, Spandauer Str. 1, 10178 Berlin, Germany
Keywords: Internet, Autonomous System, Industry, Classification.
Abstract: The Internet consists of a network of Autonomous Systems (ASs). To understand which kind of
organizations control those ASs can help to better assess the Internet structure in terms of economic
interests and reliability. The current paper proposes a novel classification approach by combining AS-
specific data with business data from the United States Securities and Exchange Commission. Furthermore,
more detailed industry classes than in previous works are considered, inspired by the North American
Industry Classification System (NAICS). Using our methodology on a recent data set, we were able to
classify 56.69 % of the considered ASs into industries. This lays a foundation for our future work on
investigating the important players of the Internet backbone as well as their economic interests and risks.
1 INTRODUCTION
The Internet expanded rapidly during the last
decade. From 2001 to 2013, the fraction of the world
population using the Internet increased from 8.0 %
(International Telecommunication Union, 2011) to
an estimated 38.8 % (International
Telecommunication Union, 2013), with a
simultaneous population increase from 6.1 billion
(United Nations Population Fund, 2001) to 7.2
billion people (U.S. Department of State, 2013),
resulting in approximately 2.8 billion Internet users
today versus 0.49 billion in 2001. This rapid growth
in users resulted in a heterogeneous and complex
system, making analysis and modelling of the
Internet structure difficult.
Our paper is part of an on-going research project
that is investigating how the Internet of today is
structured in terms of economic interests, control
and reliability. Who are the important players of the
Internet backbone, what are their economic interests
and risks with respect to their business models, and
what are the implications for reliability, security and
privacy as well as political control?
Our first step towards approaching these goals is to
classify the important organisations that control
Autonomous Systems (ASs) of the Internet
according to business categories, which could
support future analyses along all of those
dimensions. For example, with respect to reliability
and security, common methods assess the robustness
of the Internet structure based on graphs and
modelling the Internet as an abstract complex
network consisting of nodes (each representing an
AS) that are connected via edges. However, such
approaches solely focus on topology-based
robustness and so far ignore the highly economically
driven character of the Internet, as well as
corresponding heterogeneous risks of attack and
control.
At an organizational and global routing level of
abstraction, the Internet can be considered as
composed of ASs. An AS can be defined as “a group
of IP networks run by one or more network
operators with a single clearly defined routing
policy; when exchanging routing information to the
outside, each AS is identified by a unique number
(Réseaux IP Européens, 2011). The Internet
Corporation for Assigned Names and Numbers
(ICANN) and, via delegation, the Regional Internet
Registries (RIR) are responsible for registration of
these AS numbers (ASNs). The amount of registered
ASs increased from roughly 10,000 in the year 2000
to more than 60,000 in 2013 (Potaroo, 2012), which
is also another indicator for the substantial increase
of Internet complexity.
Classifying the major players of the Internet
backbone is an interesting challenge in itself because
publicly available business data is sparse. Our
approach presented in this article focuses on
361
Baumann A. and Fabian B..
Who Runs the Internet? - Classifying Autonomous Systems into Industries.
DOI: 10.5220/0004936803610368
In Proceedings of the 10th International Conference on Web Information Systems and Technologies (WEBIST-2014), pages 361-368
ISBN: 978-989-758-023-9
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
analysing the public registration information for AS
numbers. Moreover, we present an approach for the
classification of ASs into detailed industry classes in
order to better understand the organizational and
economic patterns of the Internet.
The rest of the paper is structured as follows:
Section 2 discusses related work. Section 3 presents
the data sources, followed by Section 4 on our
methodology. Section 5 presents our results, and
Section 6 concludes the paper.
2 RELATED WORK
Some earlier research articles proposed approaches
for classifying ASs into various categories. The
classification approach used in our paper was
initially inspired by the methods employed by
Dimitropoulos et al. (2005). Based on an expert
system that uses text classification techniques, the
authors used organization names to categorize ASs.
Each AS was assigned to one or more of the basic
classes Internet service providers (ISP), Internet
exchange points (IXP), network information centers
(NIC), companies providing no Internet service as
well as education- and research-, military-,
government- and health-related networks. The
authors were able to classify 20,598 out of 32,689
ASs in 2005, which corresponds to 63.01%.
Another work (Dimitropoulos et al., 2006) used
even more coarse-grained classification categories,
namely only large and small ISPs, customer ASs,
universities, IXPs and NICs. The method applied
was based on the AdaBoost algorithm (Freund and
Schapire, 1997) using several attributes (e.g.,
organization description; number of inferred
providers, customers and peers; number of
advertised IP prefixes) to classify the relevant ASs
into their respective classes. The authors were able
to classify 95.3% of 19,537 ASs with an accuracy of
78.1%.
The main focus of the work by Chang et al.
(2005) was to estimate traffic volume between
individual ASs. For this, the authors classified ASs
regarding their initial utility, which resulted into the
three classes web hosting, residential access and
business access. The methodology used by the
authors is different from other work conducted in
this area. Instead of investigating an individual AS
and assigning it to a class, they created a class and
tried to find relevant ASs on the Internet. The
authors were able to identify 56% of all BGP-
advertised ASs with their approach.
The primary focus of the paper by Dhamdhere and
Dovrolis (2011) was to analyse the evolution of the
AS ecosystem over the last 12 years. ASs were
classified into the classes enterprise customers, small
and large transit providers, access/hosting providers
and content providers. A decision tree approach was
applied for classification. In order to build the
training set, for each class 50 ASs were classified
manually. Afterwards, the classification was
conducted for 42,000 ASs by using the number of
customers and the number of peers as independent
variables. Classification accuracy for the classes
ranged between 76% and 82%.
All of those articles have in common that the
proposed classes are not comprehensive and do not
resemble real industries. Thus they contribute not
much to a better understanding of the industry
structure behind the ASs comprising the Internet.
Our work addresses this research gap by proposing a
classification approach that adopts fine-grained
industry classes.
3 DATA SOURCES
3.1 CAIDA
The Cooperative Association for Internet Data
Analysis (CAIDA) “is a collaborative undertaking
among organizations in the commercial,
government, and research sectors aimed at
promoting greater cooperation in the engineering
and maintenance of a robust, scalable global Internet
infrastructure.” (CAIDA, 2011). One project offered
by CAIDA is the AS Rank project (CAIDA, 2012). It
is based on Border Gateway Protocol (BGP) routing
data collected by RouteViews (2013) and the RIPE
NCC (2013). The list of ASs that is used in our
paper contains the information of 59,576 ASs. An
excerpt of the dataset can be seen in Figure 1. For
the purpose of classifying ASs into industry classes
mainly the org name attribute was considered as
highly relevant.
Figure 1: Excerpt of CAIDA AS Rank.
3.2 SEC
The U.S. Securities and Exchange Commission
(SEC) is a government agency in the USA (United
States Securities and Exchange Commission, 2013).
Its primary purpose is to regulate securities and
WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies
362
enforce federal securities laws. Every company
publicly traded in the United States has to file
certain documents with the SEC. The Electronic
Data Gathering, Analysis, and Retrieval (EDGAR)
system makes those filings available to the public.
This can be used to gather the Standard Industrial
Classification (SIC) code for the company (Figure
2). An SIC code can be directly mapped to an
NAICS code using a mapping table (CareerOneStop
U.S., 2013). Thus it is possible to uniquely identify
the industry of an AS’s organization by use of the
EDGAR system. A limitation is that only
organizations that are listed on a stock exchange in
the USA can be found in the system.
Figure 2: Excerpt of SEC EDGAR result.
3.3 RIR as Information
As an additional information source, data from the
RIRs was retrieved. The website cidr-report.org
contains AS information from all RIRs. It allows
searching for individual ASs and returns the
information that comes from the WHOIS services of
the individual RIRs. In order to simplify the data
retrieval process, this website was also used to
retrieve AS-specific WHOIS information instead of
using the WHOIS services of different RIRs. A
sample of such information can be seen in Figure 3.
Figure 3: Sample of RIR AS information.
4 METHODOLOGY
Figure 4 gives an overview of the process of
classifying the ASs presented in this paper. As a first
step, the relevant industry classes for the
classification approach needed to be defined. Their
definition draws from the North American Industry
Classification System (NAICS, 2013). Due to the
intrinsic online setting of our investigation, special
adjustment was necessary, meaning that several of
these classes were either merged, dropped or
changed. In the case of ASs, some industries are
missing at all while some of them are
overrepresented. Therefore, the NAICS was only
used as a basis for the classification approach in our
particular setting. An overview of the classes can be
found in the Appendix.
Figure 4: Process of AS industry classification.
Step 1: Preprocessing. The initial AS list included
data from the year 2012 and was taken from the
CAIDA AS Rank project (CAIDA, 2012); it
contained 59,576 ASs. In order to only include
reasonable and recent data, the list was
preprocessed. At first, the information gathered from
the RIRs was used to filter for inactive ASs. This
reduced the list by 17,830 ASs, leaving 41,746 ASs
to classify. Furthermore, all ASs that did not have an
according organization name, i.e., all entries either
containing no specification of the underlying
organization or being a no registry entry, were
removed from the list. Eliminating 1,362 ASs, this
step left 40,384 ASs in the list.
Step 2: Keyword Classification. In the next step, a
keyword list was created by analysing word and
phrase frequencies with the help of an occurrence
counting of words, bi-grams and tri-grams. All
words and phrases that appeared quite frequently
were analysed in more detail. It was assumed that
tri-grams needed to occur at least five times, bi-
grams ten times and simple words twenty times to be
selected for deeper analysis. The rationale behind
AS list from CAIDA
AS Rank Project
Definition of
Industry Classes
Step 1:
Preprocessing
ASs to be Classified
Step 2:
Keyword
Classication
Keyword
Classification
Result
Keyword List
Step 3:
SEC
Classication
SIC Codes
SEC Edgar
Final Classification
Result
WhoRunstheInternet?-ClassifyingAutonomousSystemsintoIndustries
363
this procedure was to include only those words and
phrases that are most frequent and therefore
important. This makes it possible to classify several
ASs at the same time based on a single phrase or
keyword. Keywords were mainly defined in such a
way that the organization name or a part of it had to
comply with the complete keyword. This means that
for example in case of the keyword “ship” only the
word itself would fit and not “membership” or
“ownership”. This was done to ensure the reliability
of keywords by avoiding undesired mismatches. The
selection of keywords itself was randomly cross-
checked based on real data to further ensure their
reliability and unambiguity. Only those words or
phrases were chosen whose unambiguity in relation
to industry classification was satisfactory. For
example the keyword “Internet service provider” is
highly reliable if it comes to sorting into the
category ISPs & Networks, while “service provider”
might lead to wrong results for the same category.
Organizations having a (part of their) name such as
“content service provider” would also fit into such a
category.
In order to minimize wrong categorizations, an
iterative learning process was applied. The
procedure was as follows: based on the first
selection of keywords, the AS numbers were
categorized into the industry classes created so far.
Each category was then checked for wrong
categorizations. For this purpose, the list of
categorized ASs and their underlying organization
was reviewed manually. If the categorization of an
ASN was wrong, the reason was identified and
eliminated with the help of refined or discarded
keywords. This procedure helped to ensure that only
those keywords remained that are at the same time
reliable and general. In order to check for further yet
not identified keywords, a list was generated that
contained all non-categorized ASNs. This list was
then manually checked for further keywords at each
iteration. This was particularly important in case of
misspelling and language-specific variations. For
example, the keyword “university” was represented
by many language specific variations such as
“universitas”, “universidad” or “univ”. An example
for misspelling is “network infomation center”
which occurred at least seven times in the list. Such
variations were additionally included in the keyword
list for each category.
Based on this extended and refined keyword list,
the procedure started from the beginning and was
repeated again. The complete list of the industry
classes created and their respective definition are
shown in Figure 7. The keywords used for each
industry class are given in the Appendix.
Figure 5: Ambiguous EDGAR search result for “Sprint”.
Step 3. SEC Classification. A Java program was
written to download information from the SEC
EDGAR system. The organization name was used to
search for the company. For 40,384 search requests,
2,732 entries could be found in the EDGAR system.
However, sometimes the same company has several
names, which resulted in more than one outcome for
the organization name. An example of such an
ambiguity can be found in Figure 5. Because there
was no reliable way to uniquely identify the correct
entry in such a case automatically, all entries with
multiple search results were eliminated which led to
1,706 remaining search results. Furthermore some
companies had no SIC code and were eliminated as
well. This resulted in 469 ASs that could
additionally be classified into industry groups.
5 RESULTS
5.1 Keyword Classification
Applying the method described above and using the
keywords shown in the Appendix to classify the
40,384 ASs, resulted in 22,786 or 56.42 % of
classified ASs. The industry class distribution based
on keyword classification only can be seen in Figure
6. According to this data, most frequently the
organizations that own ASs belong to the industry
classes Education & Research, Finance &
Insurance, ISPs & Networks, and Telephone &
Communications. This class distribution seems to be
intuitive: ISPs, telephone and IT companies as well
as universities have more incentives to register an
AS than for example a travel agency because ASs
classified into these categories are often related to
communications, but often also represent major
institutions that have a high tendency to own an AS
simply because of their size.
WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies
364
Figure 6: Industry class percentages based on keyword
classification.
Of all clusters (apart from the generic Company
cluster) ISPs & Networks is the category that is most
frequent and accounts in case of both classification
approaches for around ten percent of all the
classified entities. This is an expected result since
ASs pertain to the communications business, and
offering Internet access is a key factor in this
particular business area. The categories Education &
Research and Telephone & Communication occupy
the second and third positions with 5.32 % and 5.24
% respectively. Therefore, even today, companies in
the area of Internet and information technology are
still overrepresented because of their particular
Internet affinity. A bit more unexpected, however, is
that financial institutions seem well represented
since the fourth position is taken by the Finance &
Insurance cluster. All other categories are smaller
with percentage values between 2.57 %
(Government & Military) and 0.07 % (Space).
However, a limitation of our results is that the
general Company cluster still encompasses 22.54 %
of the classified ASs. This fact and the remaining
number of unclassified ASs indicate that there is still
a potential for improvement regarding the
classification process. Yet it is questionable whether
it is possible to reach much better results with semi-
automatic classification approaches because of the
presence of non self-explanatory organization names
and acronyms such as NGM or EDP. Not only is it
difficult to classify those simply based on keywords,
it is also challenging to specify what kind of
organization they represent without further manual
and individual investigations.
5.2 SEC Classification
The industry-class frequency of organizations based
on an alternative classification that is solely based
on SEC data is shown in Figure 8. The industry
classes Construction & Manufacturing as well as
Figure 7: Industry classes with their definitions and
percentages based on the keyword classification.
Consulting & Management have the most
organizations with ASs. The classes Government &
Military, ISPs & Networks, IT & Internet Service,
IXPs, NIC, as well as Space have no ASs at all.
However, because not all companies are listed with
the SEC and in particular governmental institutions
and privately held companies are not registered, the
lack of representation of these classes is inherent.
With the help of the SEC data it was possible to
classify additional 116 ASs, which could not be
classified via keywords only (Figure 9).
Furthermore, the industry classes of 206 ASs could
be specified more precisely which had previously
been assigned to the Company class (Figure 10).
Most newly specified classifications were assigned
to the Construction & Manufacturing as well as the
Consulting & Management industry classes. Their
prevalence reflects the results of the SEC
classification.
By combining both classification approaches we
were therefore able to classify 22,892 of the 40,384
AS of the preprocessed list.
This accounts for 56.69 % of all considered ASs
that could be assigned to an industry class. The final
result is shown in Figure 11.
Cluster Definition Result Ratio
Address
ASs where no underlying organization is specified but an address,
where the AS itself, the underlying organization or its managerial
unit is located.
326 0.81 %
Company
Collecting bucket for those ASs which are hard to categorize based
on their organization name but at least can be identified as a
company.
9,104 22.54 %
Construction &
Manufacturing
Mostly building firms and manufacturers are part of this class. 97 0.24 %
Consulting &
Management
ASs related to advising and leading of a company. 139 0.34 %
Education &
Research
ASs related to learning und gaining of new insights such as schools,
universities, research facilities and networks as well as laboratories.
2,150 5.32 %
Entertainment &
Information
ASs which are for example related to television, gaming, radio or
publishing.
761 1.88 %
Finance &
Insurance
This class consists mainly of banks and insurance firms. 1,664 4.12 %
Government &
Military
ASs with an authority and military character as well as areal
territories such as cities and states are relevant for this class.
1,039 2.57 %
Healthcare
Next to hospital (district) related entities, this class contains
pharmaceutical firms.
695 1.72 %
IT & Internet
Service
ASs that are affiliated with online as well as offline IT services and
computer products. In general, this includes those firms which
provide a service or product that is based on the Internet or IT, but
which do not offer Internet access.
1,371 3.39 %
ISPs & Networks
Collects ASs of those organizations which offer Internet access or
provide the necessary infrastructure.
3,775 9.35 %
IXPs
This class collects all ASs which function as exchange point in the
Internet.
134 0.33 %
NIC
Contains those ASs which are “responsible for managing and
allocating Internet resources” [6].
390 0.97 %
Space Contains ASs of the area of astronautics. 30 0.07 %
Telephone &
Communication
This class contains (mobile) telephone providers and sellers as well
as general communication-based organizations.
2,118 5.24 %
Travel
Contains all ASs that are related to mobility and travel, such as
airports, train stations, hotels and travel agencies.
103 0.26 %
Trade & Transport
Collects ASs of the area of wholesale and logistics including apparel
and food.
232 0.57 %
Utilities
Organizations which provide electric power, water as well as other
basic materials; also services such as waste disposal, coal and mining
belong to this class.
256 0.63 %
WhoRunstheInternet?-ClassifyingAutonomousSystemsintoIndustries
365
Figure 8: Industry class sizes based on SEC classification.
Figure 9: New classifications.
Figure 10: More precise SEC classification.
6 CONCLUSIONS
This paper proposed a classification approach for
categorizing ASs into detailed industry classes in
order to better understand the economic background
of the Internet structure. The industry classes are
inspired by the NAICS (2013), which had the effect
that an unprecedented level of detail regarding the
industry classes for classification could be achieved.
Data was mainly obtained from the CAIDA AS
Rank project as well as from SEC.
The classification of ASs into industry classes
based on their underlying organization revealed an
on-going strong dominance of telecommunication
Figure 11: Final classification using both keyword and
SEC data
(Note: It was possible to categorize an AS into
more than one industry class.).
and IT-related firms in the current Internet as well as
of large institutions such as banks and universities.
It was possible to classify 56.69 % of all ASs (after
preprocessing). Nevertheless, the amount of
unclassified ASs indicates that there is room for
improvement regarding the categorization process.
A refined and extended keyword selection process
could provide better results. Nevertheless, since
there is a non-negligible amount of ASs having
organizational specifications that are not self-
explanatory or acronyms, this would involve a
difficult challenge.
Some of our further explorative attempts to find
new ways for AS classification with the help of
clustering algorithms had limited success so far.
However, another possible route could be to apply
methods from Natural Language Processing (NLP)
to the AS data and also for analysing search results
from the Web for acronyms or other challenging
organization names.
Moreover, customers of the various ISPs cannot
be captured by the current method. It is often the
case that large Internet providers also represent
smaller customers who are not registered in the
organizational information of the ASs. Here,
studying the level of IP addresses could provide
further insights but will also involve complex
challenges.
Cluster Result Ratio
Address 326 0.81%
Company 8,898 22.03%
Construction & Manufacturing 198 0.49%
Consulting & Management 254 0.63%
Education & Research 2,155 5.34%
Entertainment & Information 771 1.91%
Finance & Insurance 1,685 4.17%
Government & Military 1,039 2.57%
Healthcare 698 1.73%
IT & Internet Service 1,371 3.39%
ISPs & Networks 3,775 9.35%
IXPs 134 0.33%
NIC 390 0.97%
Space 30 0.07%
Telephone & Communication 2,134 5.28%
Travel 104 0.26%
Trade & Transport 2,546 6.30%
Utilities 269 0.67%
WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies
366
Various other classification approaches might be
feasible. In future work we will try to find other
valuable classification systems aiming to take an
even closer look at the composition of the Internet.
Furthermore, we will use our classification results to
further investigate the important players of the
Internet backbone as well as to assess their
economic interests and risks, at individual as well as
global scales. Moreover, we aim to derive
implications for Internet reliability and control
assessments as well as for security and privacy
analyses.
ACKNOWLEDGEMENTS
The authors thank Sebastian Dombrowski for his
programming work during parts of this research.
REFERENCES
CAIDA, 2011. About CAIDA. http://www.caida.org/
home/about/. (Access Dec, 2013).
CAIDA, 2012. AS Rank Project. http://as-
rank.caida.org/?mode0=as-dump-info. (Access Dec,
2013).
CareerOneStop U.S. Department of Labor, Employment
and Training Administration, 2013. NAICS-SIC
Cross/Reference.
http://www.acinet.org/industry/Ind_Sic.aspx?id=&nod
eid=1. (Access Dec, 2013).
Chang, H., Jamin, S., Mao, Z., Willinger, W., 2005. An
Empirical Approach to Modeling Inter-AS Traffic
Matrices. In Proceedings of the 5th ACM SIGCOMM
Conference on Internet Measurement.
Dhamdhere, A., Dovrolis, C., 2011. Twelve Years in the
Evolution of the Internet Ecosystem. IEEE/ACM
Transactions on Networking 19(5):1420–1433.
Dimitropoulos, X., Krioukov, D., Riley, G., claffy, k,
2005. Classifying the Types of Autonomous Systems
in the Internet. SIGCOMM 2005 Poster, Philadelphia,
Pennsylvania.
Dimitropoulos, X., Krioukov, D., Riley, G., claffy, k.
2006. Revealing the Autonomous System Taxonomy:
The Machine Learning Approach. In Passive and
Active Network Measurement Workshop (PAM),
Adelaide, Australia.
Freund, Y., Schapire, R.E., 1997. A Decision Theoretic
Generalization of Online Learning and an Application
to Boosting. Journal of Computer and System Sciences
55(1):119–139.
International Telecommunication Union, 2011. Internet
Users per 100 Inhabitants 2001-2011. http://
www.itu.int/ITU-D/ict/statistics/material/excel/2011/
Internet_users_01-11.xls. (Access Dec, 2013).
International Telecommunication Union, 2013. Key ICT
Indicators for Developed and Developing Countries
and the World. http://www.itu.int/en/ITU-D/
Statistics/Documents/statistics/2013/ITU_Key_2005-
2013_ICT_data.xls. (Access Dec, 2013).
NAICS, 2013. North American Industry Classification
System. http://www.census.gov/eos/www/naics/.
(Access Dec, 2013).
Potaroo, 2012. 32-bit AS Number Report. http://
www.potaroo.net/tools/asn32/. (Access Dec, 2013).
RouteViews, 2013. University of Oregon Route Views
Project. http://www.routeviews.org/. (Access Dec,
2013).
Réseaux IP Européens, 2011. AS Number Assignment
Policies. http://www.ripe.net/ripe/docs/ripe-525.
(Access Dec, 2013).
RIPE NCC, 2013. Réseaux IP Européens Network
Coordination Centre. http://www.ripe.net/. (Access
Dec, 2013).
U.S. Department of State, 2013. World Population Day
2013. http://www.state.gov/secretary/remarks/
2013/07/211828.htm. (Access Dec, 2013).
United Nations Population Fund (UNFPA), 2001. The
State of World Population 2001 – Demographic,
Social and Economic Indicators. https://
www.unfpa.org/swp/2001/english/indicators/indicator
s2.html. (Access Dec, 2013).
United States Securities and Exchange Commission
(SEC), 2013. The Investor’s Advocate: How the SEC
Protects Investors, Maintains Market Integrity, and
Facilitates Capital Formation. http://www.sec.gov/
about/whatwedo.shtml. (Access Dec, 2013).
WhoRunstheInternet?-ClassifyingAutonomousSystemsintoIndustries
367
APPENDIX
Cluster K eywords
Address avenue, building, flat, floor, strasse, gpo box, handelsweg, mcpo box, no., po box, road, street, suite(s), tower
Company associates, agency, a\.s\., bv, b\.v\., cjsc, co, co kg, companies, company, coporation, corp, corporation, d.o.o., de
c.v., enterprise(s), gmbh, inc, incorporated, l\.l\.c, limited, llc, llp, lp, l\.p\., ltd(a), organization, s.a. de c.v., s.p.a.,
s.r.l., sp. z o.o., srl, s\.a\., s\.l\., sa, sas, sl, trust, z\.s\.p\.o, zspo
Construction &
Manufacturing
architect(s), builders, building company, building society, construcoes, construction, constructora, constructors,
electronics, machine, manufacturer(s), manufacturing, producers
Consulting &
Management
beratung, business solutions, capgemini, consultance, consultancy, consultant(s), consulting, ernst & young,
management company, pricewaterhousecoopers
Education &
Research
.*universitaet, academic, academisch, colegio, college(s), desire2learn, ecole, education(al), fachhochschule,
forschungsgemeinschaft, forschungsgesellschaft, fraunhofer, institute, instituto, knowledge network, laboratories,
laboratory, labs, learning, mitre, physics, polytechnic, recherché, research, school(s), science(s), supercomputer,
supercomputing, univ, universidad, universidade, universitaet, universitaria, universitas, universite, universiteit,
universitesi, universitet, universiti, universities, university, univerzitet
Entertainment &
Information
advertising, bbc, bertelsman, book(s), broadcasting, entertainment, football, fun, game, gaming, library,
magazine(s), mcgraw-hill, media, marketing, medien, multimedia, news, newspaper(s), printing, publications,
publishing, radio, reuters, television, times, tv, weather, zdf
Finance &
Insurance
allianz, american express, asset management, assurance, banca, banco, bank, banka, banque, blue shield, capital,
credit, finance, e\*trade, financial, goldman, guggenheim, hsbc, insurance, investment, leasing, payment, real
estate, reinsurance, rental, societe generale, stock exchange, stonepeak, visa
Government &
Military
administration of, agency, air force, army, authority, board of, bureau of, city of(fice), committee, commonwealth
of, congress(ional), council, county of(fice), department of, dept, district of(fice), dod, embassy, federated states,
gov, government, house of, iles de, military, ministry, nato, navy, northrop grumman, parliament(ary), province
of, senate, state of, united nations, united states postal service, US geological survey
Healthcare bayer, blood, dental, drug(s), drugstore, elektromedizinische, emergency, health(care), hospital(s), johnson &
Johnson, klinikum, medical, medicine, medizinische, merck, novartis, pfizer, pfizerswitzerland, pharma(cy),
pharmaceuticals, pharmafarm, propharma, social security, transplant
IT & Internet
Ser vi ce
akamai, apple inc, computer hardware, computer products, computer science, computer service(s), computer
software, computer solutions, computer systems, content provider, content service provider, content solution(s),
data center(s), data corporation, data processing, data service(s), data solution(s), data systems, dell, fujitsu,
general electric, google, hewlett-packard, host, hosting, ibm, information systems, information technology,
internet service(s), internet systems consortium, it services, microsoft, neterra, network service(s), network
systems, oracle, othello, samsung, sap, schuberg philis, siemens, sony, sungard availability, thinktech, verisign,
web service(s), yahoo
ISPs & Networks aol, arcor, at&t, backbone, broadband, bt italia, cable network(s), cogent, comcast, connection(s), esnet, exatel,
fibernet, freenet, gts, iletisim hizmetleri, internet access, internet provider, internet service provider(s), internet
solution(s), isp, lattelekom-apollo, level 3, linxtelecom, netassist, netcologne, network access, network provider,
network service(s), network solution(s), networks, ntt america, optical network, prometey, qwest
communication(s), reseau national, reseau regional, retn, road runner, rostelecom, singtel optus, smartcity, sprint,
surfnet, swisscom, t-2, telecom, telekom, telia latvija, teo, time warner cable, towerstream, transit, true internet,
uzbektelecom, verizon, versatel, vimpelcom, west call, wireless
IXPs exchange point, internet exchange, internet exchange point, ix, ixp(s), link, open exchange, peering exchange
NIC afnic, american registry, apnic, arin, east-ukrainian, internic, network information center, network infomation
center, network information centre, nic, ripe ncc,
Telephone &
Communication
alcatel, bell canada, communication(s), e-plus, elisa, ericsson, lambdarail, mobile, motorola, nokia, o2, phone,
radiotelephone, rockefeller group, singtel optus, telecommunication(s), telecomunicaciones, telefonica,
telekommunikation, telekomunikacije, telekomunikacja, telekomunikasi, telecomunicazioni, telephone(s),
telianet, turkcell, vodafone
Trade &
Transport
amazon, apparel, clothing, coca-cola, fedex, food(s), logistic(s), logisticare, retail(ers), shaya magazacilik,
shipping, shoe(s), supply, trade, trading, transport(ation), wal-mart, wholsesale
Travel air canada, airline(s), airport, bahn(hof), boing, flughafen, klm, lufthansa, hotel(s), railway, reisebuero, resort,
travel, vacation
Space aeronautics, aerospace, astronomy, nasa, space administration, space agency, space research, space telescope
Utilities bp, coal, electric power, electricity, energy, farmer, farms, fiber, gas, mine, mining, offshore, petroleum, utilities,
utility, waste, water
WEBIST2014-InternationalConferenceonWebInformationSystemsandTechnologies
368