Establishing
Stable Internet Connection For Field Work In REMOTE AREAS OF
Jaime R. Marso
Department of Earth Sciences
jmarso@geology110.com
Updates coming in December 2007 when Remote testing of Sprints EVDO Broadband Internet Service will be combined with Acrobat Connect for live audio/video transmission from Fossil Canyon, Imperial, California via the internet to an audience of students from Grossmont College in the Distance Learning Course, Geology 110 - section 5515. This broadcast will be a virtual live field trip from within Fossil Canyon on December 1, 2007.The broadcast will be bi-directional and recorded for review. More Info on Updated work
ABstract
The purpose of this
study is to provide follow-up information to the research completed in the
Miocene Tectonostratigraphy and Basin Geometry of the Central and
Background
In the event of an earthquake like the one affecting
The advancements in technology surrounding mobile/wireless
and satellite data transmission could prove to be beneficial to geologic
research in the field. The feasibility
of implementing these new technologies, in the remote locations throughout the
Java region was questionable but worthy of investigation. A generous grant from ChevronTexaco allowed a
team of geologists to travel from
Figure 1: Global location of area of interest.
INTRODUCTION
An internet connection while in the field is a valuable and
time saving tool. It helps to optimize field
time while allowing communication with associates at a different location or
even in a different country. It allows
data to be processed directly from the field.
After the field work was completed in Central and
In order to understand the optimization techniques outlined
in this paper for improving internet connectivity and speed, it is important to
understand the technical problems encountered during the initial study. GeoMapper,
ArcView, ER Mapper, and Stereonet software packages on Fujitsu tablet
computers, along with digital imaging, were used during the 2005 research
project in Central and East Java (see figure 2 for regional project location
map in Java) to illustrate the plate-tectonic history of the area. The success
of the project relied heavily on the integration of software applications and
hardware devices. The process for
achieving this cooperative interplay between devices and programs, although
devised in the lab, had to be synchronized in the field.
Figure 2: Processed
Satellite Image of the East Java Study Region
Semarang Image 457,
modified.
Every day after the field
data had been collected, analysis of the data continued back at base camp. Corel Draw 9 was used to create the stratigraphic
section models. Digital photos were
edited and captioned. GPS and geological
measurements were recorded in GeoMapper and ArcView. Geological maps were created and
updated. Rock samples taken in the field
were sent to a lab in
It was this research that
presented the perfect opportunity for experimental field trials using GIS. GIS, in combination with the wireless
transfer of data via satellite and infrared connectivity, was explored as a
means to successfully transfer data from the field to a home-based computer in
real-time. It was through GIS that geographic data could be captured and stored
while being able to update, manipulate, and display the information for
analysis. The field research would
benefit from the ability to access data in real-time. Having satellite imaging available to verify
outcrop positions and formation lithologies while in the field would save
valuable research time and improves accuracy.
Additionally, real-time transference of data would allow faster analysis
providing the field worker with valuable feedback to determine if an area
warranted more in-depth examination.
After gathering field data for the East Java Project,
research teams were restricted by the delays in data processing and needed to
wait hours, if not days, to achieve the desired results. During this project it was discovered that
establishing a network communication infrastructure while in an austere
environment carried with it some inherent problems. A team member had to sometimes carry the GPS
unit high onto a nearby hill in an attempt to establish a signal.
Equipment Used in the 2005
Project
During the 2005 project, a Fujitsu Stylistic tablet PC was
used, which was equipped with ArcView9.0 and GeoMapper software. Data were transferred from the Fujitsu to an
IBM ThinkPad-T20 using Windows-supported Infrared Data Association (IrDA). IrDA made it possible to establish a
connection between the devices but required that they be within one meter of
each other and pointed at specific angles to allow proper alignment between the
two computers.
This required a balancing act because the two computers were
not designed to have their respective IrDA sensors aligned with each
other. After the field data were
transferred to the IBM, the IBM was connected by a Motorola 9505 satellite
phone to the Iridium satellite system for wireless Internet transmission to
Other equipment included an Olympus C-2100 digital camera
for taking photos in the field and a Sony DCR-TRV30 HandyCam for creating small
video files to transfer. A small device called a Universal Serial Bus (USB) 2.0
Multi-Function Flash Card Reader proved an effective way to quickly transfer
data from the flash memory cards though a USB cable to the IBM ThinkPad
computer. This USB device supported a
12Mb/s transfer rate and is ‘hot swappable’, meaning the device can be plugged
in or removed without turning off the connecting machines. The flash card reader was useful for the
retrieval of photo images from the digital camera and for transferring data
from the Sony HandyCam, since both devices had removable flash memory
cards. The device was small, light, and
portable allowing for data to be transferred quickly and efficiently in the
field.
Working with multiple hardware devices and software
applications that are not always compatible presents a number of
challenges. “The first challenge to
implementing wireless mobile GIS is the short
communications range of wireless networks and the requirement for broad
bandwidth communication” (Tsou, 2004, p.162).
Establishing a system that provides the ability to transfer large data
packets across untried connection systems involves some elements of trial and
error. The SDSU team established a
real-time networking system for the secure transfer of data and communication
prior to departing for field work in
In addition, before addressing the challenge of accessing
the Internet to transfer data, information had to be successfully transferred
from the Fujitsu computer used in the field to the IBM ThinkPad computer used
to connect to the Internet.
Connection Speed and
Availability
Limited options for transmitting and receiving data existed
in the remote areas of
When not using the satellite system for transferring data
from the field, a dial-up Internet service provider (ISP) was used. Transfer speeds and accessibility were major
obstacles to overcome. Often field work
continued until well after dark and by the time team members returned to the
hotel, public access to Internet was closed.
When access to public Internet was available, the transfer speed made
transferring images and large files extremely slow. Occasionally a hotel was equipped to connect
to the Internet though their local phone company, Telkom. However, they generally had only one phone
line. In Purwodadi, arrangements could
be made to use their phone line for data transferring only during the hours
between
Ideally, a broadband connection needs to be in place to send
and receive data efficiently from the field.
During my research in Java, I was introduced to M. Cholil from Artha Mas
Cipta, an ISP in Java that delivers broadband service throughout
Figure 3 is a diagram to show the relationship between the
principal components of the internet network communication system, the field
geologist in and the receiving station at
Figure 3: Network Diagram of Internet Connection.
The field tests in
Figure 4: Transmission locations throughout Java with approximate connection speeds.
Results of Field Trials in 2005
Three packets of data, comprised of image
and text files, were sent from three locations while performing field research
in the
During this research project, the only way to transfer data
was to physically carry it back to the university in
To transfer an image file 450 kilobytes in size at the
Iridium ISP transfer rate of 10kb/s it should have taken six minutes. The actual time was double that (twelve to
thirteen minutes) due to the inability to maintain a constant satellite
connection and having to re-establish connections several times during a single
download. The concept of establishing an
infrastructure to exchange data remotely in real-time was explored to determine
its potential benefits and feasibility for use under actual field conditions,
and not to specifically record, monitor or otherwise determine exact download
and upload times per byte.
All of the remote areas where data transfers were completed
and resulted in similar transfer times with the exception of those attempted
from Purwodadi where it was impossible to establish a satellite connection
stable enough to complete a transfer.
The slow speed associated with connecting to the Internet via a
satellite phone is a known factor. The
purpose of the data transfers from the field in 2005 was to see if the
connection could be established and maintained in this remote area long enough
to transfer packets of data, speeds ranged of 11 to 15 kb/s. The actual transfer times differ from the
estimated times because of dropped signals occurring during the transmissions
(see Table1).
Table 1. Data Transfer Rates
In Central and Eastern
|
Field
Location:
|
File
Size(s): (kb)
|
Connection |
Estimated
Transfer Time: (s)
|
Actual
Transfer Time: (s)
|
|||
|
|
|
|
|
|
|||
|
Salatiga
|
258
|
144
|
13
|
19.8
|
11.0
|
64.0
|
50.0
|
|
Cepu
|
258
|
144
|
11
|
23.5
|
13.5
|
71.0
|
44.0
|
The internet connection speeds achieved during this project
varied depending on location. Some
connections were established within the cities and some from the field as noted
in Figure 3. Some cities had small internet
cafes, and others allowed connection only by special arrangement and only when
data transfers would not affect normal use of the phone line for business
purposes.
IDEntying the Problems of the 2005 Project
The goals of the Central and
In January 2006, additional testing was performed from
remote desert locations in southern
Figure 4:
Transmission Locations in
Testing times for data transfer from the remote locations in
southern
OPTIMIZED METHODS FOR REMOTE CONNECTIONS
The field tests performed in the southern
is the possibility of a stand-alone satellite dish and supporting internet service provider. This type of system would require advanced set-up and expense not normally available to the field geologist. This follow-up study focused on the development of a cost effective system that requires minimal technical background to establish the internet connection for field work from the remote setting.
The field tests in southern
Both connection speeds and actual transfer times for tests
conducted in
Table 2: Data
Transfer Rates for the Indonesian and
|
Field Location:
|
File
Size(s): (kb)
|
Connection |
Estimated
Transfer Time: (s)
|
Actual Transfer Time: (s)
|
|||
|
|
|
|
|
|
|||
|
Salatiga
|
258
|
144
|
13
|
19.8
|
11.0
|
64.0
|
50.0
|
|
Cepu
|
258
|
144
|
11
|
23.5
|
13.5
|
71.0
|
44.0
|
|
|
|
|
|
|
|||
|
|
258
|
144
|
15
|
17.2
|
9.6
|
31.5
|
27.2
|
|
|
258
|
144
|
14
|
18.4
|
10.3
|
51.0
|
38.5
|
Close review of the results of the 2005 Central and
phone at the field location and do not require ground lines or supporting infrastructure at or around the remote location. This is an important design element to allow for uninterrupted connection even during local disruption (see figure 3).
The testing, both in southern
A : The software used by the
Iridium phone and network, attempts to compress and "accelerate" the
connection by performing optimizations to a subset of known internet protocols
such as Hypertext Transfer Protocol (HTTP), Post Office Protocol (POP), Simple
Mail Transfer Protocol (SMTP), and File Transfer Protocol (FTP). The software used in both the actual field
collection in
B: Long latency networks such as the Iridium network can suffer performance issues with TCP enabled applications. TCP network connections are highly sensitive to packet loss and connections should be
optimized for the target network to ensure maximum throughput without excessive retry transmit times in the event of packet loss. Optimizing the TCP window size could greatly improve throughput by addressing issues with "TCP Slow Start"
C: If information needs to be encrypted
for transmission, the encryption process should be done ahead of time using a
standard encryption method such as Pretty Good Privacy (PGP) and not done
"on the fly" over the connection, as this can add to latency issues when
recovering from packet loss. It should
also be noted that a standard file that is encrypted and transferred over the
satellite link over a support "accelerated" protocol such as FTP
would benefit from possible protocol optimizations from the Iridium client, the
connection would not benefit from any compression since the encrypted
information would have extremely low compressibility due to the randomness of
the encrypted information. Any
compression should be done before the data is encrypted. A program like PGP
could save time when the data being transferred was of a sensitive nature and
required encryption in the field. Where
cost is a factor in designing the networking system, it should be noted that
cost free, open-source, software is available to encrypt data (GnuPG, 2007).
Summary and CONCLUSIONS:
The study shows that it is possible to use a satellite based
internet link in remote geologic areas where land-line internet service is not
available or services has been disrupted by a natural disaster. There are drawbacks for designing a relatively
inexpensive connection, namely limits on bandwidth. This will affect the size of the files and
the time it takes for them to be sent and received. Technology is in a constant state of change. It is important to realize that advancements
in one area require adjustments in another.
The equipment used to design this simple and workable solution to
internet connectivity in
Possible optimized transmission steps:
0: Ensure TCP window size is
optimized for low latency high packet
loss
connection.
1: Isolate field data for
transfer
2: Compress and package field
data for transmission. (Winzip,
Gnuzip, Etc)
3: Encrypt compressed and
packaged field data. (PGP, GnuPG, Etc)
4: Establish satellite connection
using accelerated and optimized
connection
software on computer provided by phone manufacturer.
5: Transmit data over accelerated
TCP protocol per the satellite
carriers recommendation, such as HTTP, FTP, SSH. Note: this varies by satellite carrier.
In conclusion, a more robust and resilient communication system can be developed to establish a remote internet connection that would benefit geologic work in the field and provide valuable communication from remote settings. A system to communicate from remote areas is a simple process using existing technology and equipment. It does not require advanced technical skills and can provide reliable access to the internet. Its success will depend on following the steps to optimize the data for transmission by optimizing the TCP/IP window size for communication on the host operating system, pre-compressing and encrypting the dataset, and using a standardized method of communication over the satellite network that can be optimized. Previous results from the 2005 Central and East Java project along with subsequent testing in the southern California desert regions were reviewed by Kevin Workman, Senior Staff Engineer, at Qualcomm, and a determination was reached that it would be reasonable to expect a 50% increase in connectivity speed if the steps in this study are implemented (K. Workman, personal communication, April 20, 2007).
In the future it is recommended to use a standard protocol
that has compression and acceleration enabled over the satellite network. In the case of Iridium that would be
In 2006,
ACKNOWLEDGEMENTS
The original research project in East and
I would also like to acknowledge the teamwork between SDSU and the University of Gadjah Mada (UGM), without which this project would not have been successful. I would like to give a special thank you to Wahyu Wilopo in the Geological Engineering Department at UGM, for his help and guidance, and to Dr. Subagyo and Dr. Sudarno, who both spent hours in the field instructing me. I would like to thank Dr. Eric Frost, Department of Geological Sciences, SDSU, for his efforts in organizing and problem solving all project-related issues, as well as his personal help and guidance in processing and interpreting the data gathered during my field research in Java.
I’d like to offer special recognition and thanks to Mr.
Kevin Workman of Qualcomm for his collaboration in evaluating the results of
the
REFERENCES
GnuPG Org. (2007), retrieved
Iridium Data (2004), Technical Advisory, Options for
Connecting USB Computers to an ISU, Rev. 1.0, Iridium LLC,
Iridium Satellite, LLC, (2007), Retrieved
Marso, Jaime R. (2005), Miocene
Tectonostratigraphy and Basin Geometry of the Central and East Java Region, San
Diego State University, pp. 38-46.
Tsou, Ming-Hsiang. (2004).
Integrated