Docstoc

baker

Document Sample
baker Powered By Docstoc
					HPWREN article final
                                                                              Page 1 of 17

THE HIGH PERFORMANCE WIRELESS RESEARCH AND EDUCATION NETWORK:

                                    AN OVERVIEW

                                 Jeff Baker, UC San Diego

Funded by the National Science Foundation’s Research Experience for Undergraduates Program



       Many people enjoy the convenience of having a wireless device, whether laptop,

PDA, or other remote system. While wireless is proving to be a significant convenience

in the realms of research, technology, and daily consumer life, it has become the primary

reason that residents in the most remote areas of San Diego County can now access the

Internet, even with the absence of electricity. The existence of such a wireless network

enabled over a wide area magnifies the possibilities that wireless can achieve, especially

in the wake of rapidly increasing high-speed communication needs (Figure 1).

       Such a wireless environment is being created by University of California–San

Diego Principal Investigator (PI) Hans-Werner Braun, Co-PI Frank Vernon, and their

team of collaborators throughout San Diego County. Funded by the National Science

Foundation’s Advanced Networking Infrastructure and Research (ANIR), Braun and

Vernon continue to build out the High Performance Wireless Research and Education

Network (HPWREN), to provide high-speed Internet access to multiple field science and

remote education locations in San Diego County. Astronomers, ecologists, geophysicists,

educators, and students all benefit from HPWREN connectivity while Braun, Vernon,

and their research team continue to explore further possibilities of network expansion.
HPWREN article final
                       Page 2 of 17
HPWREN article final
                                                                               Page 3 of 17

       Figure 1. The 802.11b standard provides a balance between range and signal strength

necessary for such a network to operate at maximum efficiency, where the original 802.11

standard provided for higher connection speeds but limited range. The use of this standard

enables 45 megabits-per-second (Mbps) backbone node connectivity to be possible – an integral

part of HPWREN’s 2,500 square-mile, wide-area network architecture.
HPWREN article final
                                                                             Page 4 of 17

WIRELESS TECHNOLOGY: THE INS AND OUTS

       Wireless technology has been accessible for several years, applicable in almost

every facet of society today; hence, wireless network research is not a new discussion

topic. However, HPWREN is unique with respect to application and development,

making its mark on wireless research with varied implementation and utilization, and

acting as a testbed for network infrastructure measurement and analysis. This high-speed

network facilitates numerous applications in remote and hard-to-reach places, utilizing

unlicensed microwave spectrums. Backbone connections operating on the 5.8 Gigahertz

(GHz) spectrum function at an amazing 45 megabits per second (Mbps).

       As a wide-area network, or WAN, the HPWREN architecture is designed with

particular objectives in mind, which include developing a network to provide high speed

information connectivity in areas where such access is not otherwise possible. HPWREN

pushes the envelope in wireless point-to-point connectivity, with access points (including

several Native American reservations and two fire stations) operating on the license-free

2.4 GHz. These remote access points operate on the 802.11b IEEE standard for wireless

connectivity using Lucent Outdoor Router radios. Meanwhile, the HPWREN backbone

operates at 45 Mbps, utilizing Western Multiplex Tsunami radios in the 5.8 GHz

unlicensed frequency spectrum. The primary challenge in operating at these frequency

bands is that they require clear line-of-sight paths to function properly. As a result,

network nodes are placed on high elevation mountain peaks, allowing higher vantage

points for direct point-to-point links.

       The HPWREN backbone consists of high bandwidth wireless network bridges

between mountain peak nodes, such as Mount Woodson, Mount Soledad, North Peak,
HPWREN article final
                                                                                  Page 5 of 17

Toro Peak, Red Mountain, Mount Laguna, and Mount Palomar. By way of these wireless

bridges, network bandwidth is relayed to various remote locations, such as Palomar

Observatory and the La Jolla Native American Reservation. As electricity may not

always be available in these remote areas, some HPWREN sites sustain power by way of

solar panels, which provide the power for each microwave antenna to operate. Each

solar-powered 802.11b-based relay system (Figures 2 and 3) consists of four 80-watt

solar panels; four independent 94-AH gel cell batteries; a charge controller; and a DC-DC

converter. Each system is capable of generating a peak solar power of 320 watts, which is

able to continuously power a device consuming around 32 watts; this generously suffices

for HPWREN sites, which generally operate at 13.5 watts.




       Figures 2 and 3. Since many HPWREN antennae are set up in various rugged locations

throughout northeastern San Diego County, electricity may not always be readily available. Here,
HPWREN article final
                                                                                   Page 6 of 17

the La Jolla Native American reservation relay antenna can operate up to five days on solar-

charged batteries alone, and will continue to operate–even through variable weather conditions.>
HPWREN article final
                                                                              Page 7 of 17

NETWORK RESEARCH: MEASUREMENT AND ANALYSIS

         Technology used by the HPWREN researchers includes radio transmitters,

receivers, routers, and a multitude of other equipment needed to operate a wireless

network. Data collected from this hardware set is stored in a Management Information

Base (MIB), which resides locally on each piece of equipment. This data includes

received signal levels (RSL) from the Western Multiplex Tsunami radios, local/remote

signal-to-noise ratios (SNR) from Lucent radios, bit error rates (BER), router and switch

interface data, and power data gathered from various uninterruptible power supplies

(UPS).

         MIB data is a set of variables that define what information is available for any

particular machine in the network; this data also acts as a standard for each particular

component. For example, a router’s MIB interface data might include a value of total

packets that have passed through it, or the voltage at which it is operating. As such, MIB

data varies depending on the device from which it was gathered. Collected by way of the

simple network management protocol (SNMP), MIB data reveals the state of component

health as well as operational characteristics of network equipment. MIB data also plays a

role in network operations, where researchers have the ability to set variables to remotely

control equipment.

         Other measurement data collected include Active Measurement Data (AMP),

which includes round-trip time (RTT) and traceroute testing, throughput testing, and

Mping testing; and Passive Measurement and Analysis (PMA) data, which is gathered

from dedicated passive measurement machines connected to the network. These Passive

and Active Measurement machines, or PAMs, collect AMP and PMA data at various
HPWREN article final
                                                                                Page 8 of 17

locations in the network. While MIB data provides data for a specific component on the

network, PAMs offer more detailed traffic information, such as what types of packets are

being sent or received, maximum throughput, RTT measurements, and/or passive packet

traces.    In addition, PAMs allow researchers to test the limits and capabilities of

HPWREN as a whole.           PAMs can be placed at any point in the network (though

HPWREN researchers have traditionally placed them at backbone nodes).

          In addition to the aforementioned, aggregate statistics of multiple flows are also

collected from a central router, known as Netflow data (Figure 4). Netflow data provides

a summary of traffic flow through a Cisco 3660 switch located at the Mount Woodson

backbone node, and summarizes what types of network activity occur. While MIB data

is limited to a particular device and indicates how many total packets it has seen, it fails

to reveal the kind of data that goes through it, Netflow data traffic patterns are identified

by port and protocol-allowing researchers to understand how HPWREN users utilize the

network as a whole. Netflow data also provides an ongoing summary of overall traffic,

allowing HPWREN researchers the ability to collect statistical data for many years.

          When analyzed together, MIB data, AMP/PMA data, Netflow data, and data from

several other sources create a clear picture for HPWREN researchers to make

adjustments to the network; any response to improvements in infrastructure or expansion

of its capabilities that arise are included and utilized, assisting researchers in

understanding network behavior. Researchers optimize the network’s flexibility and

utility based on these observational findings as a way of maximizing efficiency for an

increasingly diverse user base, furthering efforts to understanding network dynamics and

performance.
HPWREN article final
                                                                                  Page 9 of 17




        Figure 4. Network traffic yields transmission of various kinds of packets–from file

transfer protocol (FTP) data to hypertext transfer protocol (HTTP) packets, and others. Here, a

typical day of traffic reveals that port 20 (the IANA-designated port by which FTP packets are

typically transferred) was used most, as indicated by the red portion of the graph. This type of

data assists HPWREN researchers in understanding end user requirements and network changes

that impact users.
HPWREN article final
                                                                           Page 10 of 17

HPWREN APPLICATION AND UTILIZATION

       Several projects are conducted each year by numerous research organizations

worldwide, many of which are conducted in remote field settings. How, then, are the

collected data brought back to the laboratory or classroom? In years past, varied amounts

of data were recorded on tape and delivered via ground or air. However, with the

reification of high-speed data transmission from remote sites, immense strides have been

made in geological, geophysical, ecological, and astronomical education and research.

For instance, Palomar Observatory’s 48-inch Oschin telescope (Figure 5) was recently

connected to HPWREN’s 45 Mbps backbone. As a result, astronomers around the world

are now capable of receiving real-time images from the observatory’s digital camera.

This allows for faster and more innovative applications for data processing of supernova

detection, near-Earth asteroid studies, and the like.

       San Diego State University (SDSU) Mount Laguna Observatory (MLO)

astronomers utilize this real-time connectivity in much the same way.            Remote

observation through MLO is made possible by way of HPWREN, which allows more

astronomers to use the observatory’s telescopes, thus increasing and expanding MLO’s

observer and user base.       SDSU astronomers use HPWREN regularly to transfer

astronomical images and spectra to campus, to other California State University

campuses and collaborators, to their partners at the University of Illinois, and various

other collaborators around the world. Prior to HPWREN implementation, researchers

had to send this data via 56 kilobits-per-second dial-up modem. In fact, writing the data

to tape and transporting them by car to San Diego (or even Illinois) was faster than

sending them by dial-up modem!
HPWREN article final
                                                                             Page 11 of 17

        HPWREN also plays a significant part in the observatory’s formulation and

successful development of future astronomy projects. For example, SDSU scientists are

currently implementing a high-speed frame-transfer Charge Coupled Device (CCD)

camera, intended to facilitate the study of rapid variations in cataclysmic variable stars.

This new CCD camera, which will be based at MLO, will take two-megabit images every

0.10 to 1 second, and will produce approximately 100 gigabits of data per night (during a

typical on-target eight-hour observation). With HPWREN’s 45 Mbps connectivity, this

data can be disseminated to MLO astronomers and their worldwide collaborators in real-

time.




Figure 5. Coupled with the Oschin’s digital camera, HPWREN’s 45 Mbps backbone

provides researchers at the Palomar Observatory with the ability to transmit real-time

images to astronomers worldwide.
HPWREN article final
                                                                             Page 12 of 17



       HPWREN proves to be a resource not only for astronomy research, but also for

ecological purposes as well. The Santa Margarita Ecological Reserve (SMER), a 4,000-

acre research environment situated within the Santa Ana Mountains, embodies the next

generation of field science observation. Researchers from around the world are now able

to utilize HPWREN’s backbone network to collect real-time data from field instruments

located on-site at the reserve. This allows their laboratories around the globe to remotely

access data as it is being gathered at SMER (Figure 6). One capability includes the use of

cameras to make direct habitual observations possible. Acoustic sensors give SMER

scientists the ability to monitor wildlife activity, such as animal calls, and disseminate

this data immediately from the field to laboratories worldwide.
HPWREN article final
                                                                             Page 13 of 17

Figure 6. HPWREN-connectivity makes data collection and observation more efficient. SMER

field scientists remotely send and receive data directly from the field, allowing for rapid

dissemination of real-time data to laboratories worldwide.



        Just as immediate ecological data streams to laboratories, HPWREN fosters

immediate observational capability in the field of seismological research. Seismologists

today can observe the earth from the inside-out, allowing greater insight of earthquake

and other seismic activity patterns. By way of seismic sensors connected to HPWREN

(Figures 7 and 8), fault lines like the San Andreas and San Jacinto Faults are monitored

for seismic wave activity; in addition, fault structures are mapped out three-dimensionally

as data is streamed to researchers and scientists elsewhere. HPWREN co-PI Frank

Vernon, a researcher from Scripps Institute of Oceanography’s (SIO) Institute of

Geophysics and Planetary Physics (IGPP), and other field scientists are using real-time

data collection and dissemination to send and receive continuous real-time data from

remote research stations, which allows for previously inaccessible data to be collected

and distributed efficiently.
HPWREN article final
                                                                                  Page 14 of 17




Figures 7 and 8. Numerous seismic sensors coat the San Jacinto fault zone from the Buck Ridge

Fault in the south to the Hot Springs Fault at the north end. These sensors consist of Streckeisen
HPWREN article final
                                                                              Page 15 of 17

STS-2 seismometers with Reftek 24-bit broadband digitizers linked to the central multiplexing

and relay site on Toro Peak, located in the Santa Rosa Mountains.

        Throughout all the field research and network analysis, the spirit behind

HPWREN remains: education, both for HPWREN researchers analyzing the network,

and for Native Americans utilizing the wireless network in San Diego County. Soon all

18 of San Diego County’s Native American reservation learning centers will have access

to broadband via HPWREN, where many rural centers like these customarily access the

Internet via dial-up modem. The HPWREN access is a collaboration between UCSD

researchers and the recently formed Southern California Tribal Chairmen's Association

(SCTCA) Tribal Digital Village Network (TDVNet), which is funded by Hewlett Packard.

        The first of these remote learning center HPWREN connections involved the Pala

Native American Reservation, where over 600 tribal members (including over 150

elementary school students) reside. Before the learning center was linked to HPWREN,

the students were only able to access the Internet via 24kbps dial-up modem; now, they

are able to do research for their homework assignments in a more timely manner–thanks

to HPWREN. Additional learning centers connected to HPWREN include the Rincon

and La Jolla Native American reservations–both of which are also located in the rural

area of northeast San Diego County.
HPWREN article final
                                                                            Page 16 of 17

CONCLUSION

       Researchers and collaborators from throughout San Diego County all benefit from

the ease of use and increased efficiency that HPWREN has provided thus far. As a

means to further goals, the HPWREN team continues to pursue future expansions of the

network.      Research and development of wireless infrastructure excels within the

HPWREN environment by existing as an active testbed for network research–reifying the

utility of such a network in both the consumer realm and the educational scene, and

allowing researchers and educators to efficiently process data in the fields of astronomy,

geophysics, ecology, computer science, and seismic activity.

       As additional opportunities for expansion and development arise, further strides in

network research and development continue for the HPWREN team and their

collaborators. For more information about HPWREN news, facts, figures, or projects,

visit http://hpwren.ucsd.edu.



ACKNOWLEDGEMENTS


       The author is a participant in the Research Experience for Undergraduates

Program, which is funded by the National Science Foundation. Specifically, the author is

affiliated with the High Performance Wireless Research and Education (HPWREN)

project, which is sponsored by the National Science Foundation and its ANIR division

under Grant Number ANI-0087344.         HPWREN is based on work by the National

Laboratory for Applied Network Research (NLANR) Measurement and Operations

Analysis Team (MOAT) project, which is sponsored by the National Science Foundation

and its ANIR division under Cooperative Agreement Number ANI-9807479.
HPWREN article final
                                                                           Page 17 of 17

       The HPWREN team is creating, demonstrating, and evaluating a non-commercial,

prototypical, and high-performance, wide-area, wireless network in San Diego County.

The network consists of backbone and access sites transmitting over the unlicensed

frequencies of 2.4 and 5.8 GHz respectively. The high performance backbone network

(5.8 GHz) is comprised of 45Mbps duplex point-to-point links. The access points (2.4

GHz) are located in hard to reach areas and are comprised of point-to-point and point-to-

multipoint links.


       Any opinions, findings and conclusions, or recommendations expressed here are

those of the author and do not necessarily reflect the views of the National Science

Foundation.


       The author would like to give special thanks to HPWREN Principal Investigator

Hans-Werner Braun, Computer Scientist Todd Hansen, and Senior Writer Kimberly

Mann Bruch for their assistance with this article.




BIOGRAPHY


       Jeff Baker is an undergraduate student at the University of California, San Diego.

He is currently pursuing his bachelor’s degree in Philosophy, and has been employed in

computer science-related positions for almost 4 years.

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:6
posted:11/15/2011
language:English
pages:17