
April 2004
DLT in Latin America
Speaker: Hikmet Sari, Supelec, France
Dates: 18 - 30 August 2003
Cities: Lima, Peru;, Bogota and Bucaramanga, Colombia; Quito,
Ecuador;
and Panama City, Panama
It all started at the Welcome Reception of ICC 2003 in Anchorage
last May when Bhaskar Sengupta, Director, Membership Programs
Development, asked me if I would be ready to go on a Distinguished
Lecture Tour (DLT) in Latin America. I presume he had had positive
feedback from the Asia-Pacific Region, where I went for a two-week
tour covering Taipei, Hong Kong, Seoul, Tokyo, and Singapore in
November/December 2002, and wanted me to go to another region. Since
my only trips to Latin America in the past were to Rio de Janeiro and
San Salvador, I wanted to see more of this region, and I gave a
positive answer to Bhaskar. He immediately introduced me to Jorge
Hedderwick, Director, Latin America Region. I then had a short
conversation with Jorge and volunteered to go on a tour in Latin
America provided there was sufficient interest from the Local
Chapters in my topics and we could find a convenient time period. I
indicated to him that the best time would be August, just before the
term starts in our schools and universities.
I had almost forgotten that conversation in Alaska when I received
an email from Jorge on 4 July confirming that there was strong
interest from Peru, Colombia, Ecuador, and Panama, and asking me to
inquire about flights. In terms of timing, it was just perfect: I was
invited to give a series of lectures during the last two weeks of
August, just between my summer vacation and the beginning of the
term. Jorge informed me that Ms. Araceli Garcia Gomez, Guadalajara
Chapter Chair, would be organizing my tour working with the Local
Chapters in the host countries. Araceli did a great job making the
contacts and arranging the logistics. The tour was organized over two
weeks so I had two days of lecture in each country and one day for
traveling from one country to another.
In my previous tour in the Asia-Pacific Region, I only gave one talk
in each visited country, and the topic was broadband wireless access,
which is my primary talk advertised on the ComSoc Website. For this
tour in Latin America, there was a bigger expectation from the Local
Chapters, and I was asked to give two or more lectures on different
topics in each country. Fortunately, in addition to my usual lecture
on broadband wireless access, I had a presentation ready on broadband
cable access. With the exception of Colombia, where I was asked to
give the first lecture in two different cities, I gave the two
lectures in each country and thus covered the fields of both wireless
and cable access.
The tour started on Monday, 18 August, with a flight from Paris to
Lima via Miami. Every time I have an overseas flight with a
connection, I get worried about missing it. The flight to Miami was
perfectly on time, and I immediately proceeded to the check-in
counter of the airline for the flight to Lima. I had a real fright
when the lady at the counter looked at my ticket and said "Sir, your
reservation to Lima is for Wednesday, 20 August." Suddenly, my talks
in Lima scheduled for the next two days seemed in jeopardy. I was
preparing myself to tell her that I needed to get to Lima whatever it
cost when she unexpectedly tore my ticket and handed a boarding pass
to me. It seemed unbelievable, because sometimes when something goes
wrong, everything goes wrong. But I was lucky this time; I had just
been given a boarding pass to get to Lima that same day and at no
additional charge. I felt so relieved, and thought, always double
check and even triple check the work of your travel agent and
assistant!
On my arrival at the Lima airport, Mr. Jefferson Lopez Goycochea
from the University of San Martin de Porres was waiting for me. My
two talks in Lima were scheduled as invited talks within the
framework of INTERCON 2003 (the 10th edition of an international
conference on electrical and systems engineering) organized by the
Student Branch of that university. I gave my talk on broadband cable
access on Tuesday morning, and my talk on broadband wireless access
on Wednesday afternoon. The attendance was about 40 people. One of
the reasons for not having more people was perhaps that the talks
were a part of a large conference with a number of parallel sessions.
The talks were translated into Spanish by a student working with me
in ping pong mode. Every half a minute, I would stop to give him the
floor to translate what I had just said.
My short stay in Lima was memorable. The Local Chapter had some
other guests from Brazil, Argentina, and France, and they organized a
tour for us in the old city on Wednesday morning. We visited the
places of interest of Lima, and I learned quite a bit on the history
of Peru. I will also remember the restaurant Los Cantaros, where we
had two lunches and I discovered the delicious Seviche and Pisco
Sour. You should try these specialties in Los Cantaros whenever you
have a chance to visit Lima.
My flight from Lima to Bogota on Thursday, 21 August, was at 6 a.m.,
so that night was very short for me. Fortunately, I was able to get
some sleep during the flight and was in a good shape when I arrived
in Bogota. There, Aldo Forero, Chair of the Colombia Chapter, was
waiting for me. After lunch and some rest at the hotel, Jose David
Cely, Treasurer of the Chapter, took me to Distrital University to
visit the local Student Branch. I was impressed by the number of
students who came to talk to me. After the visit, a group of students
took Jose-David and me for dinner, and we continued to exchange our
views on different topics. The group was so friendly!
My talk in Bogota was scheduled for Friday, 22 August, 9 a.m. at the
University of San Buenaventura. When I arrived at the university, I
saw a big line of people waiting to register for the lecture.
Obviously, the lecture was strongly advertised. At 9 a.m., the
amphitheater was completely full. I was told that the main
amphitheater had a capacity of 160 seats only and that, since over
270 people registered, videoconferencing facilities were installed so
that the remaining people could attend on a large screen in the
amphitheater next door. I was honored when Mr. Frey Fernando Garzon,
Rector of the university, welcomed me in both Spanish and French, and
most of all, when they played the French national anthem, La
Marseillaise, for me. This was the first time in my life that had
been done!
In the afternoon, Jose David took me to the airport and we traveled
together to Bucaramanga, where my next lecture was scheduled. During
the flight from Bogota to Bucaramanga the sight from the airplane
window was truly impressive; even more impressive was the landing at
the 2000-m-high airport at the top of the mountains. When we arrived
at the hotel, a group of students from the local Student Branch were
waiting to take us for a short city tour and dinner. On this Friday
evening the city was very lively, but still seemed very peaceful.
My lecture was scheduled for 9 a.m. on Saturday morning. The
attendance was around 200, mostly students from local universities.
This is a huge number for a mid-size provincial city like
Bucaramanga. After my talk, I was interviewed by representatives of
the local Student Branch. They asked me a long list of questions on
my view of the future of telecommunications in Latin America, the
possibilities of local research, and so on. I tried to advise them as
much as I could.
In the afternoon, Jose David and I flew back to Bogota, and the next
day (Sunday, 24 August) I left Bogota for Quito, the capital city of
Ecuador. At the airport, I was welcomed (in perfect French) by Galo
Gascante Lopez, the local Section Chair. Taking advantage of my
visit, the chapter had organized a two-day seminar entitled
"Broadband Access: State of the Art and Future Trends." My lectures
were scheduled as the morning sessions of the seminar, which had some
50 attendees. The Monday afternoon session was a presentation
entitled "Broadband Services" from the national operator Andinatel.
The Tuesday afternoon session consisted of a presentation followed by
a demo on ADSL technology from the local Alcatel organization. In
this way, the seminar covered all broadband access technologies
including wireless, hybrid fiber/coax cable, and ADSL. No translation
into Spanish was provided for my two talks, but the afternoon
sessions by local speakers were in Spanish. The audience was from
industry, service providers, and universities. This seminar was a
good example of how a local ComSoc chapter can combine a DLT with
other presentations and organize a larger event.
Quito is almost on the equator. Galo took me to the monument erected
on the equator at longitude 78° 27' 8" west, which is a major
attraction for both visitors and locals. There, you can have your
picture taken with one foot in the northern hemisphere and one in the
southern hemisphere, which we both did. The next day we visited the
historic part of Quito, where I admired the numerous churches and
other colonial buildings. The old city is truly worth seeing,
particularly after dark with its magic illuminations.
On Wednesday, 27 August, I had my flight from Quito to Panama City,
the final stop of my DLT. At the airport, I was welcomed by Gustavo
Giaz, Chair of the local ComSoc Section, and Eduardo Gonzalez,
Treasurer. After checking in at the hotel, I was taken to the office
of the local IEEE chapter, where the Chair, Tania Quiel, was waiting
for us. After signing the Guest Book and having a few pictures taken
in the office, we all went for dinner in a splendid seafood
restaurant.
My lectures in Panama City were scheduled as late afternoon sessions
from 6 to 9 p.m. (on Thursday and Friday, 28 and 29 August) so that
people from academia and industry and other organizations could
attend after their daily work. A professional interpreter was hired
to provide simultaneous translation of my talks into Spanish.
Attendance was quite modest (3035 people), essentially because
it was still the summer vacation period, but there was a lot of
interest in the lectures, judging from the number of questions and
the discussions after the talks.
During my stay in Panama, Gustavo took me on a tour along the Panama
Canal and a city tour. At the canal, we visited the main operations
center, where you can still see a lot of machinery from the 1930s.
Panama City is an interesting city that gathers the characteristics
of Latin America and U.S. cities.
Overall, this tour was very rewarding. On one hand, I was happy to
serve the local chapters and help student branches in several
countries of Latin America. On the other hand, I discovered new
countries, met very friendly people, and enjoyed the Latin
hospitality. With many of them, I will be keeping in contact, and we
will surely meet again in the near future, be it in Latin America or
here in Paris.
Designing an Advanced Regional Research Network
Arturo Azcorra, Ignacio Soto, Iván Vidal, Carlos
García,
and Francisco Valera (Universidad Carlos III de Madrid);
Rafael Sánchez (Nortel Networks); and Angel Pérez
(Telefónica de España)
Background
The Spanish authorities are fostering the development of high-speed
regional research networks as a better way to guarantee that adequate
infrastructures can be built at the regional level to access the
Spanish and international research networks. In this context, the
government of the region of Madrid imposed on itself the challenge of
creating a high-speed regional research network to be one of the most
advanced within Europe. It established a collaboration with
Universidad Carlos III de Madrid in order to introduce the most
advanced technology, yet find a cost-effective solution. This has led
to the design and deployment of REDI Madrid, a regional research
network for the research and education institutions of the region of
Madrid that is under the supervision of a non-profit organization
called Fundación Madrimasd para el Conocimiento.
In this article we share our experience in setting up the
requirements and designing this advanced network.
Figure 1. Physical topology of REDI
Madrid.
|
Requirements
After a consensus process with different institutions, the design
requirements for the new REDI Madrid network were listed as:
- High performance: Provision of a minimum access speed of
1 Gb/s.
- Reliability: Institutions are day after day more
dependent on their data communications to be able to attend to their
daily activities, so high reliability was considered crucial.
- Scalability in terms of capacity: Network technology is
constantly evolving, and new and more bandwidth-demanding
network-based applications are appearing. This means that to be able
to keep pace with these developments, the network must provide easy
ways to update capacities and access speeds.
- Scalability in terms of number of connected institutions:
This would allow a phased network deployment, in which institutions
could progressively join the network.
Regarding the institutions to be connected, it was decided to opt
for a phased approach. Out of the 70 relevant research and education
institutions in the region of Madrid, nine very relevant ones (in
terms of size, bandwidth demands, and research results) were selected
for the first phase, defining at the same time a roadmap to gradually
incorporate the others. The institutions to be connected in the first
phase were Consejo Superior de Investigaciones Científicas
(CSIC), Instituto Nacional de Técnica Aeroespacial (INTA),
plus the public universities located in Madrid: Universidad de
Alcalá (UA), Universidad Autónoma de Madrid (UAM),
Universidad Carlos III de Madrid (UC3M), Universidad Complutense de
Madrid (UCM), Universidad Nacional de Educación a Distancia
(UNED), Universidad Politécnica de Madrid (UPM), and
Universidad Rey Juan Carlos (URJC).
Network Design
After studying the different available possibilities, dense
wavelength-division multiplexing (DWDM) was chosen as the core
transmission technology. DWDM allows sharing of the expensive fiber
optic infrastructure and provides a straightforward growth path. It
also allows network self-restoration when ring fiber topologies are
used.
It was decided to use DWDM equipment selected from the Metro product
family in order to reduce costs, taking into account the involved
distances. Using long-haul DWDM equipment would have resulted in two
disadvantages: first, equipment costs surpassing the savings derived
from the shared optical fiber infrastructure; and second, lack of
efficient multiservice capability support. However, it would have
made the physical fiber design easier because of greater distances
supported.
Commercial equipment from four different vendors was analyzed,
resulting in the selection of the Nortel Networks Optera Metro
5200 family. The reasons for this are that Optera provides carrier
class reliability, is well suited to the involved distances, provides
enough capacity (up to 32 protected lambdas in one ring), and is
designed for scalability (10 Gb/s in one carrier are already
available, which gives 320 Gb/s over the ring).
Another design decision was to use a ring fiber topology with
protected carriers to satisfy the high-reliability requirement. The
initial topological design, based on direct distances, was to use a
single dual-fiber ring to connect all the institutions. However, this
had to be discarded when measuring the fiber distance over the actual
physical paths of the telecom providers and the actual attenuation of
the existing fibers. After considering several topologies, a
topological design jointly performed with Telefónica (the
chosen telecom provider) was selected. This design was composed of
three dual-fiber rings that connected the CSIC premises with all the
institutions in the network. This configuration has the added
advantages, over the single ring one, of allowing more growth
capacity in each ring, and improving reliability because a failure in
one ring does not affect the remaining ones. The physical topology of
the network is presented in Figure 1.
The target topology of carriers over the rings was a logical star
that supported, for each of the connected institutions, a protected
optical carrier (lambda) to the central node (CSIC-CTI). It was then
necessary to design, for each ring, the optimal way to provide these
carriers. The carrier design implied reaching a compromise between
optical amplification and electrical regeneration on one hand, and
between initial investment in optical equipment and scalability in
terms of carriers on the other.
Optera Metro DWDM equipment uses an optical filter to extract sets
of four lambdas from the ring. The usage of the same set of lambdas
for up to four institutions in the same ring allows costs savings in
equipment (fewer optical filters), but implies that the carriers are
optically extracted at each DWDM shelf in the ring. This implies that
equipment failures, in addition to fiber cuts, would also interrupt
the ring. Additionally, extracting the carriers implies that
electrical regeneration must be performed. An alternative solution
would have been to use lambdas from different sets, passive passing
through intermediate DWDM shelves, and optical amplifiers to overcome
the attenuation (chromatic dispersion limits the fiber distance, but
this is not a problem in the considered scenario). The balance is
therefore between the cost of additional optical filters and
amplifiers minus the cost of the electrical regeneration equipment,
vs. slightly better reliability and improved scalability (cheaper
than adding a new DWDM site). The solution in this case was to place
all the carriers in the same carrier set.
The carriers that connect each institution with CSIC-CTI are
terminated by 1 Gb/s Ethernet interface cards, over which virtual LAN
and IP services are provided. A high-performance switch/
router is located at CSIC-CTI where all GbE interfaces are
connected. This equipment also provides a redundant connection to
RedIRIS (the Spanish research network) with a 2.5 Gb/s packet over
synchronous optical network (POS) interface.
Network Evolution
New institutions may be connected to the network by incorporating
additional DWDM shelves to the rings. This is not a trivial task,
since the fibers of the ring need to be rerouted to reach the newly
connected institution, and the whole ring is affected (attenuation,
equalization). An alternative approach is to deploy a point-to-point
fiber pair to one of the existing institutions and include a new
carrier at the existing DWDM shelf. It is also possible to provide
direct links between any two institutions, even if they are not in
the same ring, simply by connecting the DWDM equipment in the central
node, although an optical-to-electrical-to-optical conversion would
be needed.
It is also possible to increase the capacity for each institution.
Cards supporting one 10 Gb/s Ethernet or 10 1-Gb/s interfaces per
lambda are already available, allowing a 10x bandwidth increase.
Finally, new technologies and products that are appearing (e.g.,
allowing increased reach) will also contribute to optimizing the
optical network.
Call for Submissions
IEEE Global Communications Newsletter (GCN) seeks
original papers of general interest in the field of communications
and related areas. Topics include, but are not limited to:
- National and regional developments in communications
technologies, services, markets and standards
- Pilot experiences in communications
- Communications research and development
- Reports on national and international large-scale
projects (e.g., NSF, EU IST)
- Telecommunications convergence, regulatory, and legal matters
- Information and knowledge society
- New applications of communications in politics, health,
education, commerce, security and defense, surveillance, agriculture,
standard of life, handicapped people care, industry, tourism, space,
transportation and navigation, environment, sustained development,
globalization, and so on
- Research trends
- Market trends
- Historical perspectives in communications
- Education in communications
- Reports on key workshops or conferences
- ComSoc chapter activities
Authors willing to present research results in communications are
encouraged to avoid exhaustive or theoretical descriptions and focus
on the general interest of their work. In that case, they should cite
the sources (project URLs, journals, conference proceedings) where
detailed descriptions can be found.
Authors willing to submit reports on workshops or conferences are
especially encouraged to do so in case of IEEE-backed ones, although
GCN is open to disseminate the conclusions of any event in the
field of communications.
Please check previous issues in IEEE Communications Magazine
or contact Joang.Haro@upct.es
or javier@det.uvigo.es if
there are any questions about the suitability of a paper.
Prospective authors should prepare their manuscripts preferably in
plain ASCII or MS Word format, with a maximum length of 1200 words,
and send them to either of the submission addresses below. MS Word
files may have pictures and tables embedded (subtract 200 words per
each figure or table). Alternatively, provide them as separate files
using any standard coding. Only send screen dumps if strictly
necessary, since they will be subjected to a minimum resolution of
300 dpi in the final version.
Submission addresses:
Dr. Joan García-Haro, Editor, Joang.Haro@upct.es
Dr. Francisco Javier González-Castaño, Associate
Editor, javier@det.uvigo.es
|