 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
 |
 |
 |
 |
 |
 |
 |
 |
|
|||||||
SOCOM TILO COTS Results of Experiment # C15: On-Demand Mobility Objective: The On-Demand Mobility experiment’s objective was to demonstrate the benefits of adding specially designed middleware to existing COTS systems and components to greatly increase levels of performance (bandwidth utilization improvements) to address ever increasing demands and requirements of the WarFighter and other mission-critical programs. The demonstration showed how On-Demand Mobility middleware addresses common COTS deficiencies and provides an architectural processing solution that eliminates the gap between COTS software and DoD performance requirements. On-Demand Mobility enables WarFighter and other Homeland Defense/First Responder systems to incorporate a far greater number of existing and future mission-critical systems into currently installed COTS hardware components.
Diagram 1– Avon Park Network Test Environment The objectives for the Avon Park experiment were to utilize existing COTS network environments and to use On-Demand Mobility technology to demonstrate real-time processing performance regardless of configuration and/or network traffic load. Performance testing centered on various network paths throughout the Avon Park environment with mobile transactions that were composed of relational database operations and data payloads. Functionality testing was centered on credentialing, accountability, image transfer, and radiation detection Quantitative/Qualitative Results: Test Equipment: Servers
Clients
Network Equipment
Test Types:
Network Test Paths:
Central Measurement Criteria: Performance Results: Interval = Time Length of Transactional Test Run
Security Note: Observations and Comments: The Avon Park testing went particularly well and demonstrated the full compatibility of On-Demand Mobility technology with existing COTS equipment. The entire experiment was run on the networking resources provided on-site. All tests were conducted with the On-Demand Mobility Server running at full system (processor) saturation. The experiment demonstrated the ability of middleware architecture to enhance mobile data processing and continue to respond in real-time in the harshest of IT environments and with severely limited hardware/bandwidth resources. Waterfront Technologies conducted two primary experiments, one in performance testing and one in specific functionality testing. Performance testing results were excellent for all network configurations. The various network paths used in the experiment provided a very informative view of what processing performance can be expected for the different communication technologies. All test configurations were able to achieve real-time performance with the worst result being 3.2 transactions a second transfer rate. On-Demand Mobility technology successfully utilized the COTS network equipment and was able to consistently deliver maximum real-time performance throughout all of test configurations. The measurements of some of the performance tests were varied because testing was performed at different intervals with unknown levels of additional network traffic from other applications and experiments. This was one of the primary premises of the experiment in that On-Demand Mobility is designed to ensure some level of real-time performance even when the network is saturated and/or the bandwidth is narrow. The performance test results clearly show real-time transactional process throughput even in the most complicated of network configurations with uncontrolled levels of intervening traffic (Network Test Path N4). The N4 configuration represents the extreme level of ad hoc networking that high-value applications are going to have to endure to be successfully integrated within any critical military operations. On-Demand mobility proved that it could maintain a steady stream of transactional traffic, automatically recovering from any errors, and performing in real-time even though the test machine was over 1000 miles away attached through a variety of different network connections (802.11b, WaveRelay, satellite and Internet). The test results for the N4 configuration exceeded expectation so to ensure test result’s accuracy an additional five (5) minute test was also conducted to confirm the results. Functional testing for the N4 configuration included picture retrieval, credentialing, and situational accountability. All mobile applications responded in a range of one to two second at the Oscar Range pulling information from an On-Demand Mobility Server both in Building 77 and in Virginia. For the observers of the Oscar Range experiment the results were startling and hard to conceptualize considering the network path being used. Demonstrating the ability to retrieve information from such a long distance through such a complex networking routing path in real-time fostered many observer suggestions for possible applications and usage. It was clearly seen that On-Demand Mobility technology can instantly connect existing computer processing resources regardless of location and connection type. Some of the functional testing results performed at the Building 77 experiment site included local and remote radiation detection using an image-based sensor application. The radiation detection application (Radiation Scouttm) was able to seamlessly perform in real-time providing radiation analysis every three seconds even while conducting performance testing (processor saturation) at the same time. Several real-time monitor programs were deployed at both Building 77 and the Oscar range that were able to monitor the results of the radiation sensor and provide real-time results. This test was also done using a radiation detection application running in Virginia that was exposed to a radiation source and provided real-time results to all the real-time monitors deployed in the Avon Park experiment. A reliability and recovery test was also performed with the radiation detection application in which the server was crashed and brought back up. The application buffered all the sensor data and reconnected with the server and automatically resynchronized itself successfully. Additional functional testing in the area of real-time credentialing and field activity accountability were performed at both Building 77 and the Oscar range. These tests clearly demonstrated the power to instantly cross-reference an Identification (ID) card or asset tag with any connected data source, even a data source that was 1,000 miles away and connected by a variety of communication mediums. ID cards were barcode scanned which initiated the real-time search and retrieval of related information including the corresponding image from a relational database located in Building 77. At this point, specific action codes were assigned to the ID that allowed the ID to be associated with a specific field activity, such as CHECK-IN, HOT ZONE IN. From this, a real-time monitor was run that provided instant situational awareness of the associated activity. For example, at the Oscar Range a set of ID cards was validated and assigned to a variety of first responder activities and tracked in real-time from Building 77. This very same experiment was also performed very successfully with only a small time delay using the On-Demand Mobility server and relational database in Virginia. One of functionality tests was usage of the PictureXpress program. PictureXpress is a real-time image transfer and retrieval program. Pictures were taken from both a Motorola Symbol device and a smartphone and instantly transferred to a relational database. Real-time monitor was used to allow the images to be organized and retrieved instantaneously. Tests using 802.11b/g responded in less than a second. Images could be retrieved from a relational database across the satellite link to Virginia. The capability impressed much of our audience. 1600x1200 images were retrieved in less than 2 seconds with full encryption. The tests clearly prove that large data files such as images can be moved back and forth in real-time across any form of IP-based COTS network. Error rates were generally manageable with almost none for local wireless traffic and less than 1% for satellite traffic. WaveRelay error rates were a little more perplexing. Test results were quite varied and it was evident that errors were generally only encountered during the initiation stages of test runs. To be more specific, when the test run was started there was a 80% greater chance of error in the first few test sequences than anywhere else in the run. This was especially the case when transferring data between the WaveRelay and the Satellite link. Also, transactional speed tended to start out very slow and then build up quickly. After about 5 seconds the WaveRelay system responded quite positively and generally executed flawlessly. Additional testing is really required to pinpoint the complete operational characteristics of WaveRelay before any real conclusions can be drawn. The Avon Park test proved that On-Demand Mobility technology can simultaneously support any number of high-value network-centric applications across any type of COTS-based network. Moreover, real-time communications can be maintained without serious interruption even under the most extreme network and operating conditions. On-Demand Mobility applications can maximize the use of current resources and provide real-time connectivity all while maintaining full security and complete transactional reliability. These operational characteristics represent the mandatory performance requirements that are needed to field any type of network-centric military application. The Avon Park test confirmed that On-Demand Mobility middleware could maintain these operational characteristics in a real world implementation. |
