|
|
 |
|
Data Acquisition and Analysis Plan
The U.S. Department of Transportation (USDOT) Next Generation 9-1-1 (NG9-1-1) Proof-of-Concept (POC) demonstration will test key features and functionalities of the envisioned NG9-1-1 system. The POC will also serve as a test-bed to validate technical feasibility and acquire important data metrics useful to the greater emergency response community. The POC test-bed equipment will be deployed within the Booz Allen, Texas A&M University, and Columbia University test laboratories and at selected Public Safety Answering Points (PSAP). Data acquisition will require hardware, software, and manual data collection methods. This document describes the data acquisition system, the analysis that will be conducted, and the metrics that will be obtained.
The objective of this document is to define a comprehensive plan for acquiring and analyzing NG9-1-1 POC data. The plan will assist the NG9-1-1 Project Team and USDOT to evaluate the technical and operational success of the POC. The data acquired from the POC will serve as benchmarks for future large-scale NG9-1-1 technology deployments and further assist in the refinement of the final NG9-1-1 transition plan. The data will facilitate the USDOT, Standards Development Organizations (SDO), industry vendors, PSAP operational community, and future independent evaluators in defining measures of interest for Internet Protocol (IP)-based emergency calling. In addition, this document will identify the data acquisition system put in place to acquire these measures including the software, hardware, and resources required.
This document includes the following—
- Documentation of the design of the data acquisition system used for collecting data from call origination, call routing, call termination, and network equipment
- Specification of the commercial off-the-shelf (COTS) and open source products used to implement the data acquisition system
- Descriptions of the data collection and analysis methods for the NG9-1-1 POC
- An outline of the objectives, hypotheses, measures of interest, and data sources for each data analysis method.
This document includes the data acquisition design and data analysis plan for the NG9-1-1 POC system. The remainder of this document is organized as follows:
- Section 2—Data Acquisition System: Details the design and implementation of the data acquisition system. A variety of software, hardware, and user testing methods will be used to collect the data from the POC test-bed.
- Section 3—Data Analysis Methodology: Discusses the data analyses that will be conducted over the course of the NG9-1-1 POC. Analysis will concentrate on four primary themes concerning emergency calling: emergency call propagation and timing, emergency call availability and quality, emergency call scalability, and call taker software usability. Each of these POC analyses and their associated measures of interest are detailed in this section.
- Appendix A—Definitions: Lists of commonly used NG9-1-1 terminology and their associated definitions.
- Appendix B—Glossary: Lists acronyms used in this document.
- Appendix C—Source References: Provides a list of published documents that were referenced while developing this document.
The USDOT NG9-1-1 POC will host a variety of infrastructure simulating an IP-based emergency call system and modeling the components defined in the NG9-1-1 Architecture. To gain the most benefit from the POC, a solid data acquisition system must be put in place. During the lifespan of the project, it is imperative to acquire functional and performance data for further analysis. This will assist USDOT and the public safety community in determining the technical feasibility of various components and interfaces of the future NG9-1-1 systems. In addition, it will potentially showcase the technical challenges that still must be overcome by currently available emergency response and communication technology. This section lays out the design of the data acquisition system, discusses the various components involved, and documents the products that will be used in the POC.
Figure 2–1 presents the design of the data acquisition system.
Figure 2–1—NG 9-1-1 Data Acquisition Design
From a data acquisition perspective, the NG9-1-1 POC testbed comprises three main entities:
- NG 9-1-1 Infrastructure and Components— The main functional components of the NG9-1-1 architecture. They provide end-to-end transfer of an emergency call and implement various aspects of the call origination, call access, call routing and call termination processes. The function and performance of these architectural entities are key to a successful demonstration of the NG9-1-1 concept. Therefore, significant effort is expended in acquiring data on the operation of these entities.
- Data Acquisition Sources— Tools and resources used to generate data regarding the operation of the NG9-1-1 infrastructure and components. They serve as the catalyst for data collection as well as support the overall data acquisition process. These tools directly interact with the NG9-1-1 infrastructure and components initiating controlled system reactions and data generation. They have varying levels of complexity and may require configuration, automation, manual inspection, or verbal interaction.
- Data Acquisition Sinks —Data collection, monitoring, and aggregation tools for the NG9-1-1 POC. They interact with NG9-1-1 infrastructure and components to extract raw data for later processing and analysis. They consist of a combination of hardware, software, and human assets.
The design of the data acquisition system is based on COTS and open source equipment. This modular, standards-based design addresses the plethora of hardware, software, and networking infrastructure that make up the NG9-1-1 POC. This approach provides more flexibility for product selection but still requires POC data analysts to develop an in depth knowledge of the acquired products.
The data acquisition system follows a standard FCAPS (Fault, Configuration, Accountability, Performance, Security) approach to system management and monitoring. This model is used by most production telecommunications system providers today. Given the developmental nature of the POC, certain aspects of the FCAPS model are more applicable to the POC and therefore are more heavily emphasized. The FCAPs model for systems management and monitoring is applied to the NG91-1 POC in the following manner:
- Fault— It is imperative that the production NG9-1-1 system is deployed in a robust manner. Given the nature of emergency calling, devices and connections must have a high level of availability. For the POC, metrics will be established on the overall uptime of the NG9-1-1 components and their interconnections. However, given the budget and research focus of the POC, redundant links and components will not be procured and no failover capability will be supported. Faults in the NG9-1-1 components will be detected through standard logging procedures (Syslog) and the Simple Network Management Protocol (SNMP). The main tools used for collecting logs and determining proper function of the devices are the SysLog Server and the Network Management System discussed in the subsequent sections.
- Configuration— In a production system, the ability to remotely configure and centrally provision devices is crucial. Given the geographically disparate nature of the NG9-1-1 system, centrally managing the NG9-1-1 infrastructure is a necessity . For the POC, all efforts will be made to allow remote access to the equipment. This is especially important because development and integration is spread across Booz Allen’s Center for Network & Systems Innovations (CNSI) Laboratory ( Herndon, Virginia), Columbia University’s Internet Real-Time (IRT) Laboratory ( New York City, New York), Texas A&M University’s I nternet2 Technology Evaluation Center (ITEC) Laboratory ( College Station, Texas) and the selected PSAPs. While numerous enterprise products exist for remote, centralized management and provisioning, this functionality will be provided on a device-by-device basis leveraging each individual product’s capabilities. For example, terminal access will be provided for routers, and Remote Desktop Protocol (RDP) ports will be opened for the servers.
- Accountability— Accountability focuses on the personnel responsible for maintaining and monitoring infrastructure. In addition, accountability provides mechanisms to ensure contractually agreed upon performance (i.e., bandwidth of 1.44 megabits per second [Mbps] or availability of 99.999 percent uptime) for acquired products or leased connections. Given the scope of the POC, a single person will be responsible for ensuring appropriate operation of the POC equipment. To assist in this task, the Network Management System will be used to remotely monitor uptime of components and leased network connections.
- Performance— Performance is one of the main focuses of the POC. To determine the technical feasibility of the NG9-1-1 system, performance metrics must be obtained. For the POC, performance is characterized according to call propagation and timing, call availability and quality, call scalability, and software usability. These metrics are both quantitative and qualitative in nature. To assist in the acquisition of these metrics, a variety of tools will be used. To generate realistic call scenarios and traffic loads on the system, network traffic generators and call simulators will be used. The specifics of these generators are discussed in Section 2.3. Upon injecting traffic into the system, network protocol analyzers and software log files will provide raw data for analysis and acquisition of performance metrics. The Project Team will collect the data manually and with the assistance of software, and conduct qualitative surveys with the call takers to determine software usability. The overall data analysis process is detailed in the Section 3.
- Security— Security is important to any production system to ensure appropriate access control, data rights, resource management, and data privacy. Even though the POC is an open, research-oriented initiative, best practices have been put in place to ensure a secure operating environment for the POC. These include creating a virtual private network (VPN) using tunneling, placing firewalls with appropriate access policy at the ingress/egress points of the network, and using strong username and password protection for all servers deployed in the POC. While stronger, more centralized measures should be used for the production system, these were deemed unnecessary for the POC. Although outside the scope of this effort, the effects of encryption on the performance of IP-based emergency networks should be investigated.
The follow sections contain details of the required products for the data acquisition system. Each section corresponds to one of the data acquisition entities previously depicted in the Data Acquisition Design, Figure 2–1.
Figure 2–2—NTP Server Architecture for the NG 9-1-1 POC
Network Time Protocol (NTP) Server
Product: AT&T Stratum 1 NTP Server
Data Acquisition Type: Source
Description: Stratum 1 NTP servers are computers attached directly to atomic (cesium, rubidium) clocks, Global Positioning System (GPS) clocks, or other radio clocks known as Stratum 0 devices. Normally, they act as servers for timing requests from Stratum 2 devices via NTP. Stratum 1 NTP servers are referred to as time servers and can usually maintain time to within 10 milliseconds (1/100 s) over the public Internet. For the NG9-1-1 POC, all infrastructure devices will be NTP compliant and contact a Stratum 1 NTP Server hosted by AT&T for their timing reference.
Acquirable Data Metrics/System Properties: Synchronized timing across the NG9-1-1 POC Test Environment. Time synchronization is key in tracing events through the POC system using software logs as well as for acquisition of accurate call propagation and timing metrics.
Figure 2–3—Network Management System
Product: OpenNMS™
Data Acquisition Type: Sink
Description: OpenNMS is an enterprise-grade network management platform developed under the open source model. It consists of a community supported open-source project as well as a commercial services, training, and support organization. OpenNMS operates on a scalable platform and provides a software implementation of the FCAPS network management model. For the POC, its main purpose will be for monitoring faults and tracking interface service (e.g., Accountability). Using SNMP, OpenNMS will acquire the availability metrics of the NG9-1-1 components and interfaces.
Acquirable Data Metrics/System Properties: Component Availability/Uptime, Network Connectivity Availability/Uptime Additional Information: http://www.opennms.com
Call Origination Software Event Logs
Product: Firsthand Technology - SIPc Software
Data Acquisition Type: Sink
Description: SIPc is a voice over IP (VoIP) software client. SIPc relies on the Session Initiation Protocol ( SIP) to establish IP-based telephony services. SIPc supports voice, video, and data media streams between SIP User Agents. SIPc stores software events to the sipc.log file. Events are documented with a textual description and corresponding timestamp.
Acquirable Data Metrics/System Properties: SIP Call Propagation and Timing Metrics, SIPc Error Events
Additional Information: http://www.cs.columbia.edu/irt/sipc
SIP Border Gateway and ESRP Software Event Logs
Product: Firsthand Technology - SIPd Software
Data Acquisition Type: Sink
Description: SIPd serves as the foundation for the SIP Border Gateways and Emergency Services Routing Proxies (ESRP). SIPd is a fully compliant, SIP Proxy Server that handles call session initiations and terminations between SIP User Agents. SIPd stores software events similarly to SIPc with a timestamp and corresponding description; however, it stores the information in a custom MySQL database.
Acquirable Data Metrics/System Properties: SIP Call Propagation and Timing Metrics, SIPd Error Events
Additional Information: http://www.cs.columbia.edu/irt/cinema
PSAP Automatic Call Distribution (ACD) Software Event Logs
Product: PSAPd Software
Data Acquisition Type: Sink
Description: PSAPd is a fully compliant, back-to-back SIP User Agent that handles automatic call distribution at a PSAP. PSAPd stores software events in a custom MySQL database with a timestamp and textual description of the event.
Acquirable Data Metrics/System Properties: SIP Call Propagation and Timing Metrics, SIPd Error Events
System Event Log Aggregator
Product: OpenNMS
Data Acquisition Type: Sink
Description: Event logging for OpenNMS is conducted with three distinct mechanisms: service polling, receipt of unsolicited messages (usually SNMP traps), and threshold evaluated against performance data. OpenNMS also provides a comprehensive “Service Collector Interface” leveraging SNMP, HTTP, and NSClient. These interfaces gather data that can then be used in performance graphs, thresholds, and network latency calculations.
Acquirable Data Metrics/System Properties: Network Performance Management, Fault Management, and Security Additional Information: http://www.opennms.org
SIP Call Simulator
Figure 2-4—SIPp Screenshot
Product: SIPStone—SIPp
Data Acquisition Type: Source
Description: SIPp is an open source test tool and traffic generator for the SIP protocol. It establishes and releases multiple calls with the INVITE and BYE methods. It can also read custom eXtensible Markup Language (XML) scenario files describing simple or complex call flows. It generates information regarding call rate, round trip delay, and message statistics, as well as supports comma-separated statistics dumps, Transmission Control Protocol (TCP)/User Datagram Protocol (UDP) over multiple sockets, and dynamically adjustable call rates.
Acquirable Data Metrics/System Properties: Call Rate, Round Trip Delay, SIP Message Statistics
Additional Information: http://sipp.sourceforge.net/
Network Traffic Generator
Figure 2–5—IxChariot Console Screenshot
Product: Ixia—IxChariot™
Data Acquisition Type: Source
Description: IxChariot is a test tool and IP traffic generator. IxChariot can emulate real-world applications and protocols to predict system performance under realistic load conditions. IxChariot consists of Performance Endpoints and the IxChariot Console. Using scripted traffic scenarios, Performance Endpoints send data to one another across the system under test (SUT). Upon completion, system metrics are acquired and presented to a data analyst at the IxChariot Console. IxChariot can support tens of thousands of connections representing hundreds of thousands of end users. IxChariot is configurable to generate traffic patterns that simulate converged IP services (voice, video, data) using industry standard protocols (IPv4, TCP/UDP, RTP, VoIP, IP Multipcast). For the POC, IxChariot will be used to obtain network performance data from the Call Origination Source (Booz Allen Laboratory) to the selected PSAPs.
Acquirable Data Metrics/System Properties: Network Throughput (Bandwidth), Jitter, Network Latency (Response Time), Packet Loss
Additional Information: http://www.ixiacom.com/products/ixchariot/
Network Protocol Analyzer
Figure 2–6—Wireshark Screenshot
Product: Wireshark™
Data Acquisition Type: Sink
Description: Wireshark is a graphical user interface (GUI)-based packet sniffer and protocol analyzer. A network protocol analyzer captures network packets and displays their data visually in a time sequential manner. I t is used for network troubleshooting, analysis, software and communications protocol development, and education. For the NG9-1-1 POC, Wireshark will be used to capture SIP messages passed to/from Call Origination User Agents, Border Proxies, ESRPs, IP ACDs, and Call Taker Workstations.
Acquirable Data Metrics/System Properties: SIP Protocol Analysis, SIP Call Propagation and Timing Metrics Additional Information: http://www.wireshark.org/
Acceptance of the Human Machine Interface (HMI) display by the call taker community is critical to the success of the NG9-1-1 solution. A rigorous User Acceptance Testing (UAT) methodology will be employed by the NG9-1-1 testing team to ensure that the HMI solution fully meets the defined usability requirements and standards.
Tools (i.e. Test scripts and Summary/Error Reports) to support the UAT phase will be developed to test the HMI display and acquire information about the HMI components. They also provide a complete evaluation of the HMI display functionality and usability. The data acquisition tools used during the UAT phases include,
- User Acceptance Test Cases and Scripts
- User Acceptance Summary Report
- User Acceptance Consolidated Error Report.
User Acceptance Test Scripts
These scripts are used to test the HMI display to ensure it is usable and supports the NG9-1-1 system requirements. The scripts support a data analyst in gathering user feedback regarding system functionality and usability based on the HMI display design specifications. The scripts contain scenarios and steps identified for testing of the HMI display. The scripts focus on the user’s experience with system functionality, layout of the display components, ease of use, availability of necessary information, and screen navigation.
Figure 2–7—User Acceptance Test Script
Product: Word Document
Data Acquisition Type: Sink
Description: Test scripts will be created to test all screens and applications of the HMI. The test scripts will also be developed to ensure that all requirements identified for the HMI by the NG9-1-1 Systems Design Document are addressed.
Acquirable Data Metrics / System Properties: Gathers information about the user’s experience with the system, including screen navigation, layout, access to application features, and ease of use
Additional Information: N/A
User Acceptance Summary Report
The User Acceptance Summary Report provides a description of all usability errors and issues identified by the user group during UAT. The report allows the testing team to track each identified error from its identification through resolution. The report is a collection of feedback and test results obtained during the UAT activities.
Figure 2–8—User Acceptance Summary Report
Product: Word Document
Data Acquisition Type: Sink
Description: Summary of errors identified during the UAT testing phase
Acquirable Data Metrics / System Properties: The Summary Report includes the following key fields:
- Problem—(i.e., “unintuitive layout of Call Record display,” “agency listing is not necessary on the Main HMI display,” etc.)
- Defect Summary—identifies the effect of the error on the call taker operating the HMI display (i.e., “call-takers may not need to view all of the fields currently contained in the Call Record screen, and the large number of fields prevents call takers from finding necessary information quickly,” “agency listing available directly from the Main HMI display prevents call takers from viewing other important data,” etc.)
- Screen components in which the error occurs—identifies the actual sections of the HMI display that are affected by the error
- Severity of the error—defines the severity of the error with regard to the HMI success and acceptance
- Change Control Board (CCB) Comments—identify resolution approach for the error and priority
- Error Resolution—outlines the steps needed to correct identified error within the HMI display or provides an explanation of the error will be addressed without modifying the display (i.e. training).
Additional Information: N/A
User Acceptance Consolidated Error Report
The Consolidated Error Report is a summary of findings of the HMI Display UAT. The report contains all errors that were identified during the testing period, their resolutions, and their status. The report also includes the testing methodology and the project schedule, as well as a breakdown of the data entry execution. The report is an all-inclusive summary of the UAT effort and findings, presented to the USDOT officials to evaluate the HMI display usability.
Figure 2–9—User Acceptance Consolidated Error Report
Product: Word Document
Data Acquisition Type: Sink
Description: Formal summary of errors identified during the UAT testing phase for a given period of time. Presented to USDOT for evaluation
Acquirable Data Metrics/System Properties: The report will include the following information:
- Testing Approach—identifies the testing approach and the teams conducting the testing activities. Identifies all phases of the testing process and the phases covered during the testing period.
- Test Execution—Specifies the tests and the affiliated requirements/system components that were conducted during the time frame of the report.
- Findings—Details findings of the testing phase. This will include the summary of usability, navigation, and application access, as well as other findings, based on the discovered errors.
- Recommendations—Contains a listing of recommendations to correct the errors and findings. The recommendations are summaries of CCB recommendations as well as error resolutions.
Additional Information: N/A
The NG9-1-1 POC requires a methodical and iterative approach through all levels of data collection and analysis. This section focuses on the process required to execute an effective data analysis methodology. The subsequent sections address the four main focuses of the NG9-1-1 POC:
- Emergency Call Propagation and Timing— T his subsection examines the need for calls to be routed in an efficient, time-sensitive manner through the NG9-1-1 system.
- Emergency Call Availability and Quality— Two subsections investigate NG9-1-1 Component and Interface availability and IP network performances effect on the quality of voice and video calls.
- Emergency Call Scalability —This subsection explores the NG9-1-1 system’s ability to scale on a need-driven basis with convergence of different media types (voice, video, and data). It also looks at innovative NG9-1-1 call overflow mechanisms and their ability to improve overall system performance.
- Call Taker Software Usability —This subsection discusses the usability of the HMI for the PSAP call taker software.
The data artifacts for each section include test case descriptions, data collection procedures, and data analysis templates. The test case descriptions provide a brief overview of each data collection event, a brief list of the NG9-1-1 system components required for that test case, the entrance and exit criteria, and the desired data outputs. The data collection procedures contain the detailed step-by-step process for conducting a specific data collection event. Finally, a template is provided for documenting and analyzing the NG9-1-1 POC data.
Figure 3–1—Call Propagation and Timing
Call propagation and timing is important to any emergency response system. When an emergency occurs, a matter of seconds can mean the difference between life and death. Currently, within the United States, a wireline emergency telephone call traverses from a call originator to a PSAP call taker in an average of 7 to 12 seconds.
The NG 9-1-1 architecture presents a variety of new call origination sources, including IP phones, IP wireless handheld devices, sensor systems, telematics systems, cellular wireless devices and legacy wireline support. In addition, the NG 9-1-1 concept allows many different media formats for emergency calling, including a combination of voice, video, and data (i.e. messaging/texting). As these technologies emerge and gain market acceptance, the public end user demands at least the same level of service as defined by legacy wireline calling. However, the call flow and processing of an NG 9-1-1 call requires a major change in approach and additional steps compared with those associated with legacy technologies such as wireline telephony. Therefore, it is imperative that during the NG 9-1-1 POC, call propagation and timing parameters are defined and tested.
The results of this analysis will assist a number of NG 9-1-1 stakeholders. The USDOT or other pertinent Federal agencies can further influence congressional policy for acceptable emergency response system performance; SDOs can leverage these metrics for industry conformance standards; and emergency response vendors and service providers can differentiate their product offerings from those they offer to their other markets.
In Figure 3–1 above there are a number of interrelated call propagation metrics. Table 3–1 defines these measures of interest.
Table 3–1—Call Propagation and Timing Measures of Interest
Measure of Interest |
Description |
Constraints and Relationships |
T Access |
This parameter represents the time for an emergency call to traverse an access network and arrive at an NG 9-1-1 border gateway. It spans the time frame from when a call originator initiates an emergency call to when the border gateway receives the SIP INVITE message. |
This parameter is highly dependent on the type of device used by the call originator, the type of access network used by the device, and the conversion mechanism used to convert the call signaling to SIP.
Each type of call origination source must be tested and will yield a unique representation of this parameter.
If a call traverses a shared access/network link, this parameter could be significantly affected by superfluous network traffic.
This parameter could be dissected further using knowledge of the underlying access technology. |
T LIS |
This parameter represents the round trip time for a border gateway to query and receive location information for a call originator. Upon receipt of an emergency call, a border gateway inspects the call stream for location. If no location is present in the call stream, the border gateway queries a location information server (LIS) using a unique identifier. The LIS will then respond with the location of the call originator. |
This parameter is highly dependent on the access technology used by the call origination source. The border gateway must be designed to specifically handle this type of call source.
The LIS will vary depending on type of access technology. For example—
- Wired Telephony: LIS = Automatic Location Identification Database (ALI DB)
- IP-based Telephony: LIS = VoIP Positioning Center (VPC) DB or Enterprise LIS solution
- Telematics: LIS = Telematics Third-Party Call Center DB
- Cellular: LIS = Mobile Position Center (MPC) DB
Some IP-based call originators are capable of embedding their location in the call stream. In this case, T LIS = 0.
This parameter could be dissected further for a given access and LIS technology. |
T Nat_LoST |
This parameter represents the round trip time for a border gateway to acquire a location resolution from a National Location-to-Service Translation Protocol (LoST) server. Using the location of a call originator, the border gateway queries a National LoST server. The National LoST server uses the location information (civic or geospatial) to resolve to an ESRP uniform resource identifier (URI). The border gateway then forwards the call to the appropriate ESRP. |
LoST has been implemented using a Hypertext Transfer Protocol (HTTP) interface. Therefore, T Nat_LoST is bound by the TCP time out window
T Nat_LoST < TCP timeout
Theoretically , a border gateway could forward traffic to a statically assigned ESRP, in which case,
T Nat_LoST = 0 |
T NG9-1-1 |
This parameter represents the time for an emergency call to traverse from a NG 9-1-1 border gateway to an ESRP server. It spans the time frame from when the SIP INVITE leaves the border gateway to when ESRP receives the SIP INVITE. |
This parameter is highly dependent on the bandwidth and latency of the IP transport network used to forward the emergency call.
If the call traverses a shared access/ network link, this parameter could be significantly affected by superfluous network traffic. Dedicated connectivity is recommended.
This parameter could be dissected further if the network topology of the IP transport network was understood. |
T Loc_LoST |
This parameter represents the round trip time for an ESRP to acquire a location resolution from a Local LoST server. Using the location embedded within the call stream, the ESRP queries a Local LoST server. The Local LoST Server uses the location information (civic or geospatial) to resolve to a PSAP URI. The border gateway then forwards the call to the appropriate ESRP. |
LoST has been implemented using an HTTP interface. Therefore, T Loc_LoST is bound by the TCP time out window
T Loc_LoST < TCP timeout
Theoretically , a border gateway could forward traffic to a statically assigned PSAP, in which case,
T Loc_LoST = 0 |
T Business |
This parameter represents the round trip time for an ESRP to acquire business rules from the business rules DB. The business rules DB can change the routing of an emergency call based on call stream parameters embedded within the emergency call. In addition, the business rules DB can designate supplemental and supportive data sources for the ESRP to contact. The business rules DB returns these modified routing and data source instructions to the ESRP. |
Business Rules and routing changes for emergency calls should be used sparingly. The more business rules contained within the Business Rules DB the longer the query and its resolution will take.
This parameter is system implementation specific. Time could vary substantially based on the DB implementation (i.e. permanent versus volatile memory storage and access) of the Business Rules DB. |
T Supp |
This parameter represents the round trip time for an ESRP to acquire supplemental or supportive data. A variety of supplemental and supportive data sources can exist. This information can be passed by value or reference, depending on the criticality of the information. |
This parameter is system and interface specific. Time could vary substantially based on the implementation of the external data source.
A federal mandate or the SDOs should define a maximum threshold for this parameter to ensure timely delivery of emergency calls. |
T PSAP |
This parameter represents the time for an emergency call to traverse from an ESRP to a PSAP ACD. It spans the time frame from when the SIP INVITE leaves the ESRP to when PSAP ACD responds with a SIP OK. |
This parameter is highly dependent on the bandwidth and latency of the IP transport network used to forward the emergency call.
If the call traverses a shared access/ network link this parameter could be significantly affected by superfluous network traffic. Dedicated connectivity is recommended.
This parameter could be dissected further if the network topology of the IP transport network was known. |
T Call_Taker |
This parameter represents the time for an Emergency Call to traverse from a PSAP ACD to a call taker’s workstation. It spans the time frame from when the call enters the PSAP ACD queue to when the designated PSAP call taker answers the telephone. |
This parameter is dependent on human interaction and call load at a PSAP. If there are periods of unusually high call volume or insufficient resources at the PSAP, the value of this parameter could be significant.
PSAPs should actively monitor this parameter to determine adequate resource staffing. |
T End-2-End |
This parameter designates the time it for an emergency call to traverse the whole system from call origination to call reception. It spans the time frame from when a call originator initiates an emergency call to when the Call Taker picks up the telephone at the PSAP. |
This parameter designates the summation of all other timing parameters, T End-2-End = T Access + T LIS + T Nat_LoST + T NG9-1-1+ T Loc_LoST + T Business + T Supp + T PSAP + T Call_Taker
This parameter should be within reasonable thresholds as determined by PSAP governing authority. |
Data Collection Test Cases
Test Case |
Native IP-based Call Propagation and Timing (ID – DC0001) |
Objective |
This test case will evaluate the NG9-1-1 system’s ability to initiate, propagate, and terminate an IP-based emergency call. The timing parameters discussed in Section 3.2.2 will be obtained to gauge overall system performance. |
Description |
An IP-enabled source will initiate a SIP-based emergency call into the NG9-1-1 system. The call will traverse the NG9-1-1 system and terminate at the desired PSAP and call taker. As the call traverses the system, component logs (i.e. Call Origination software, SIP Border Gateway software, ESRP software, PSAP ACD software) are generated that record the events of the call. Based on the component logs, timing parameters will be extracted through manual inspection and time difference calculations. |
Equipment |
NTP Server, SIPc Client Software/IP Phone (SIP-based), IP Telephony Border Gateway, LoST Server, LIS Server, ESRP, IP ACD, Business Rules Database, PSAP Call Taker Software |
Entrance Criteria |
All components must be present and operating according to the defined NG9-1-1 system design.
Basic network connectivity between laboratory environments must be established.
All hardware time sources must be synchronized using an NTP Server with a GPS or atomic clock source. NTP Server should be able to provide approximately 10 ms accuracy.
All system component log files should record events with millisecond accuracy. |
Exit Criteria |
The call terminates at the desired PSAP.
Ten iterations of this test case are successfully run. |
Data Outputs |
Log files are generated for each system component with appropriate timestamps. |
|
|
Test Case |
Legacy Telephony Call Propagation and Timing (ID – DC0002) |
Objective |
This test case will evaluate the NG9-1-1 system’s ability to initiate, propagate, and terminate a Wireline Time Division Multiplex (TDM)-based emergency call. The timing parameters discussed in Section 3.2.2 will be obtained to gauge overall system performance. |
Brief Description |
An analog telephone source will initiate a TDM-base emergency call into the NG9-1-1 system. The call will be converted to SIP and traverse the NG9-1-1 system, terminating at the desired PSAP and call taker. As the call traverses the system, components logs are generated that record the events of the call. Based on the component logs, timing parameters will be extracted through manual inspection and time difference calculations. |
Equipment |
NTP Server, Analog Telephone, Wireline Telephony Border Gateway, LoST Server, Simulated ALI DB, ESRP, IP ACD, Business Rules Database, PSAP Call Taker Software |
Entrance Criteria |
All components must be present and operating according to the defined NG9-1-1 System design.
Basic network connectivity between lab environments must be established.
All hardware time sources must be synchronized using an NTP Server with a GPS or atomic clock source. NTP Server should be able to provide approximately 10 ms accuracy.
All system component log files should record events with millisecond accuracy. |
Exit Criteria |
The call terminates at the desired PSAP.
Ten iterations of this test case are successfully run. |
Data Outputs |
Log files are generated for each system component with appropriate timestamps. |
|
|
Test Case |
Telematics Call Propagation and Timing (ID – DC0003) |
Objective |
This test case will evaluate the NG9-1-1 system’s ability to initiate, propagate, and terminate a telematics emergency call. The timing parameters discussed in Section 3.2.2 will be obtained to gauge overall system performance. |
Brief Description |
A telematics third-party call center will initiate an emergency call into the NG9-1-1 system. The call will be converted to SIP and traverse the NG9-1-1 system, terminating at the desired PSAP and call taker. As the call traverses the system, component logs are generated that record the events of the call. Based on the component logs, timing parameters will be extracted through manual inspection and time difference calculations. |
Equipment |
NTP Server, Telematics Third-Party Call Center Telephone, Wireline Telephony Border Gateway, LoST Server, Emergency Crash Notification DB, ESRP, IP ACD, Business Rules Database, PSAP Call Taker Software |
Entrance Criteria |
All components must be present and operating according to the defined NG9-1-1 system design.
Basic network connectivity between laboratory environments must be established.
All hardware time sources must be synchronized using an NTP Server with a GPS or atomic clock source. NTP Server should be able to provide approximately 10 ms accuracy.
All system component log files should record events with millisecond accuracy. |
Exit Criteria |
The call terminates at the desired PSAP.
Ten iterations of this test case are successfully run. |
Data Outputs |
Log files are generated for each system component with appropriate timestamps. |
Test Case |
Cellular Call Propagation and Timing (ID – DC0004) |
Objective |
This test case will evaluate the NG9-1-1 system’s ability to initiate, propagate, and terminate a cellular emergency call. The timing parameters discussed in Section 3.2.2 will be obtained to gauge overall system performance. |
Brief Description |
A cellular telephone will initiate an emergency call into the NG9-1-1 system. The call will be converted to SIP and traverse the NG9-1-1 system, terminating at the desired PSAP and call taker. As the call traverses the system, component logs are generated that record the events of the call. Based on the component logs, timing parameters will be extracted through manual inspection and time difference calculations. |
Equipment |
NTP Server, Cellular Telephone, Cellular Border Gateway, LoST Server, Simulated MPC DB, ESRP, IP ACD, Business Rules Database, PSAP Call Taker Software |
Entrance Criteria |
All components must be present and operating according to the defined NG9-1-1 system design.
Basic network connectivity between laboratory environments must be established.
All hardware time sources must be synchronized using an NTP Server with a GPS or atomic clock source. NTP Server should be able to provide approximately 10 ms accuracy.
All system component log files should record events with millisecond accuracy. |
Exit Criteria |
The call terminates at the desired PSAP.
Ten iterations of this test case are successfully run. |
Data Outputs |
Log files are generated for each system component with appropriate timestamps. |
|
|
Test Case |
SMS Call Propagation and Timing (ID – DC0005) |
Objective |
This test case will evaluate the NG9-1-1 system’s ability to initiate, propagate, and terminate an SMS emergency text message. The timing parameters discussed in Section 3.2.2 will be obtained to gauge overall system performance. |
Brief Description |
A cellular telephone will initiate an emergency text message into the NG9-1-1 system. The SMS message will be converted to a SIP message and traverse the NG9-1-1 system, terminating at the desired PSAP and call taker. As the call traverses the system, component logs will be generated that record the events of the call. Based on the component logs, timing parameters will be extracted through manual inspection and time difference calculations. |
Equipment |
NTP Server, Cellular Telephone, SMS Border Gateway, LoST Server, Simulated MPC DB, ESRP, IP ACD, Business Rules Database, PSAP Call Taker Software |
Entrance Criteria |
All components must be present and operating according to the defined NG9-1-1 system design.
Basic network connectivity between laboratory environments must be established.
All hardware time sources must be synchronized using an NTP Server with a GPS or atomic clock source. NTP Server should be able to provide approximately 10 ms accuracy.
All system component log files should record events with millisecond accuracy. |
Exit Criteria |
The call terminates at the desired PSAP.
Ten iterations of this test case are successfully run. |
Data Outputs |
Log files are generated for each system component with appropriate timestamps. |
Data Collection Procedure
Because of the similarity in the test cases, only one instance of the data collection procedure will be documented for all of the data collection test cases described above.
Test Case Name |
[Native IP Telephony | Legacy Wireline | Telematics | Cellular | SMS] Call Propagation and Timing |
Participants Names: |
|
Dates: |
|
Estimated Execution Time: |
8 hours/Test Case |
Actual Execution Time: |
|
Description: This data collection procedure will acquire the timing metrics associated with a specific call origination device. |
Measures of Interest Acquired: T Access, T LIS, T Nat_LoST, T NG9-1-1, T Loc_LoST, T Business, T PSAP, T Call_Taker |
Setup:
- Ensure power is supplied to all system components and they are currently powered on.
- Ensure network connectivity exists between all labs by performing basic PING connectivity testing.
- Ensure all systems components have a reliable connection to a single NTP Stratum 1 Server.
- Ensure granular software logging is enabled for all system components.
|
Procedure
# |
Action |
Expected Results |
Actual Results/Comments |
1 |
Launch Network Protocol Analyzer (Wireshark) on all applicable system components. (i.e., Call Origination Sources, Border Gateways, ESRPs, PSAP IP ACD, and Call Taker Workstation) |
Software should load successfully. |
|
2 |
For each component using Wireshark, apply a filter for monitoring IP, SIP, SDP, and HTTP traffic |
|
|
3 |
Initiate an Emergency Call from the Call Origination Source:
- IP = SIPc Client Software
- Legacy Wireline = Analog Phone
- Telematics = 3rd Party Call Center Phone
- Cellular = Cellular Phone
- SMS = Cellular Phone
|
|
|
4 |
Inspect logs on Call Origination or Transition Device to ensure a SIP INVITE was generated and forwarded to the Border Gateways
- IP = Laptop w/ Wireshark
- Legacy Wireline = Logs on IP Telephony Router (Cisco 2821)
- Telematics = Asterix Telephony Gateway w/ Wireshark
- Cellular = Logs of Cellular Media Gateway
- SMS = Logs on SMS Media Gateway (Pulsewan)
|
|
|
5 |
Inspect Wireshark on the [IP | Legacy Wireline | Telematics | Cellular | SMS] Border Gateway to ensure the SIP INVITE was received |
Use the time difference between when the SIP INVITE was initiated at the Call Origination Source and when the SIP INVITE was received at the respective Border Gateway to acquire T Access. |
|
6 |
Inspect Wireshark on the respective Border Gateway to ensure it acquired location for the Call Origination Source. Border Gateways will interact with the following systems for location acquisition:
- IP = LIS Server / Simulated VPC
- Legacy Wireline = Simulated ALI DB
- Telematics = N/A
- Cellular = Simulated MPC
- SMS = N/A (New System Required)
|
Note: Each Call Origination Source is associated with a different location acquisition system. Access methods/interfaces vary for each location acquisition systems. Therefore, different Request/ Response Pairs must be inspected using Wireshark to determine T LIS depending on the executed test case. |
|
7 |
Inspect Wireshark on the respective Border Gateway to ensure it successfully performed a LoST query |
Use LoST Query Request/Response Pair to determine T Nat_LoST |
|
8 |
Inspect Wireshark on the respective Border Gateway to ensure it successfully forwarded the SIP INVITE to the appropriate ESRP |
|
|
9 |
Inspect Wireshark on the ESRP to ensure the SIP INVITE was received |
Use the time difference between when the SIP INVITE was forwarded from the Border Gateway and when the SIP INVITE was received at the ESRP to acquire T NG9-1-1. |
|
10 |
Inspect Wireshark on the respective ESRP to ensure it successfully performed a LoST query |
Use LoST Query Request/Response Pair to determine T Loc_LoST. |
|
11 |
Inspect Wireshark on the respective ESRP to ensure it successfully performed a Business Rules DB query |
Use Business Rules DB Query Request/Response Pair to determine T Business. |
|
12 |
Inspect Wireshark on the ESRP to ensure it successfully forwarded the SIP INVITE to the appropriate PSAP IP ACD |
|
|
13 |
Inspect Wireshark on the PSAP IP ACD to ensure the SIP INVITE was received |
Use the time difference between when the SIP INVITE was forwarded from the ESRP and when the SIP OK message was returned to the Call Origination Source to acquire T PSAP |
|
12 |
Inspect Wireshark on the PSAP IP ACD to ensure it successfully forwarded the SIP INVITE to an available Call Taker Workstation |
|
|
13 |
Inspect Wireshark on the Call Taker Workstation to ensure the SIP INVITE was received |
Use the time difference between when the SIP INVITE was forwarded from the PSAP IP ACD and when the SIP Invite was received at the Call Taker Workstation to acquire T Call_Taker. |
|
14 |
Repeat this data collection process 10 times per test case [IP-based | Legacy Wireline | Telematics | Cellular | SMS] |
|
|
Data Analysis Template
Call Origination Source |
Iteration |
T Access (ms) |
T LIS (ms) |
T Nat_LoST (ms) |
T NG91-1 (ms) |
T Loc_LoST (ms) |
T Business (ms) |
T PSAP (ms) |
T Call_Taker (ms) |
T End-to-End (ms) |
IP-Based User Agent |
1 |
|
|
|
|
|
|
|
|
|
|
2 |
|
|
|
|
|
|
|
|
|
|
3 |
|
|
|
|
|
|
|
|
|
|
4 |
|
|
|
|
|
|
|
|
|
|
5 |
|
|
|
|
|
|
|
|
|
|
6 |
|
|
|
|
|
|
|
|
|
|
7 |
|
|
|
|
|
|
|
|
|
|
8 |
|
|
|
|
|
|
|
|
|
|
9 |
|
|
|
|
|
|
|
|
|
|
10 |
|
|
|
|
|
|
|
|
|
|
Mean |
|
|
|
|
|
|
|
|
|
|
Minimum |
|
|
|
|
|
|
|
|
|
|
Maximum |
|
|
|
|
|
|
|
|
|
|
Standard Deviation |
|
|
|
|
|
|
|
|
|
Legacy Wireline |
1 |
|
|
|
|
|
|
|
|
|
|
2 |
|
|
|
|
|
|
|
|
|
|
3 |
|
|
|
|
|
|
|
|
|
|
4 |
|
|
|
|
|
|
|
|
|
|
5 |
|
|
|
|
|
|
|
|
|
|
6 |
|
|
|
|
|
|
|
|
|
|
7 |
|
|
|
|
|
|
|
|
|
|
8 |
|
|
|
|
|
|
|
|
|
|
9 |
|
|
|
|
|
|
|
|
|
|
10 |
|
|
|
|
|
|
|
|
|
|
Mean |
|
|
|
|
|
|
|
|
|
|
Minimum |
|
|
|
|
|
|
|
|
|
|
Maximum |
|
|
|
|
|
|
|
|
|
Figure 3–2—NG 9-1-1 POC Component and Interface Availability
Proper operation of NG 9-1-1 system components and interfaces is vital to an effective emergency response service. As a mission critical system, the NG9-1-1 System must provide near 100 percent uptime. This can be accomplished effectively by building multiple levels of redundancy within the network and the system components. NG 9-1-1 system components should have backups and be hot swappable. In addition, there should be redundant network links that automatically failover in the case of failure or can be used for additional bandwidth in case of catastrophic events such as natural disasters or terrorist attacks.
When planning and implementing systems, it is important to know the availability of a single component or interface. These measures are usually acquired through statistical analysis of operational system components. Because the POC is research oriented, no efforts will be made to design a fully redundant or 100% available system. However, software (Syslog servers and Network Management software) will be used to track uptime of the NG 9-1-1 system components. This information will prove useful in gathering basic information on component availability. For a production system, these metrics could serve as starting points in determining the number of levels of redundancy needed to acquire a certain level of availability (e.g., 99.999% uptime) for the NG 9-1-1 system.
Measures of Interest
Table 3–2—Component and Interface Availability Measures of Interest
Measure of Interest |
Description |
Constraints and Relationships |
Component Availability (Downtime/year) |
Availability: 2 9’s (99%) = downtime less than 87.6 hours per year 3 9’s (99.9%) = downtime less than 8.8 hours per year 4 9’s (99.99%) = downtime less than 53 minutes per year
5 9’s (99.999%) = downtime less than 315 seconds per year |
Component Availability is heavily dependent on reliable hardware, software, and power sources |
Interface Availability (Downtime/year) |
See above |
Interface Availability is heavily dependent on robust networking hardware, software, power sources, and access medium. |
System Availability |
See above |
System Availability is dependent on component and interface availability. Since no redundancy is built into the NG9-1-1 POC all components / interface are consider in series. Therefore, system availability can be calculated by summing the downtime of all the components and interfaces. |
Data Collection Test Cases
Test Case |
Emergency Component/Interface Availability (ID – DC0006) |
Objective |
This test case will evaluate the NG9-1-1 component and interface availability. The availability parameters discussed in Section 3.3.2 will be obtained to gauge overall system performance. |
Description |
Over the course of the POC, a Syslog Server and Network Management System will be deployed to track the uptime of system components and interfaces. This is done through software logs that are harvested by the Syslog Server and through SNMP traps generated by network and server equipment for the Network Management System. Once the POC is complete, these logs will be parsed and a system component or interfaces uptime/downtime can be obtained. |
Equipment |
All NG9-1-1 system components and interfaces |
Entrance Criteria |
All components must be present and configured to log system events to their respective operating system (OS) syslog.
The Syslog Server must be installed and functioning properly.
The Syslog Server must be configured to harvest the components’ Syslogs at configured time intervals (i.e., Daily, Weekly, Monthly basis).
The Network Management System must be installed and functioning properly.
The Network Management System must be configured to poll system components and interfaces using Management Information Bases (MIBs.)
The Network Management System must be configured to collect SNMP traps from system components and interfaces. |
Exit Criteria |
Completion of the POC |
Data Outputs |
Software Log files are generated for each system component and stored within the Syslog Server.
The Network Management System proactively polls devices and creates logs using component and interface MIB tables. The Network Management System reactively stores logs of component and interface SNMP traps. |
Data Collection Procedure
Test Case Name |
Emergency Component/Interface Availability |
Participants Names: |
|
Dates: |
May – June 2008 |
Estimated Execution Time: |
6 months |
Actual Execution Time: |
|
Description: This data collection procedure will acquire the availability metrics associated with the NG9-1-1 components and interfaces. |
Measures of Interest Acquired: Component Availability, Interface Availability |
Setup:
- Ensure Power is supplied to the Syslog server and Network Management System, and they are currently powered on.
- Initially ensure network connectivity exists between the Syslog server and Network Management System, and all monitored components by performing basic PING and TRACEROUTE connectivity testing.
- Ensure all systems components have a reliable connection to a single NTP Stratum 1 Server.
- Ensure granular Syslog software logging is enabled for all system components.
- Ensure SNMP services are turned on for all system components.
|
Procedure
# |
Action |
Expected Results |
Actual Results/Comments |
1 |
Collect component and interface events using the Syslog Server and Network Management System |
|
|
2 |
Parse component log files looking for events that brought the component down or forced it to be restarted |
Component Availability metric obtained by summing the downtime across the time span of the POC. |
|
3 |
Parse interface log files looking for events that infer the interface was down |
Interface Availability metric obtained by summing the downtime across the time span of the POC. |
|
4 |
Repeat this action for each component and interface for which logs were collected. |
|
|
Data Analysis Template
|
Month 1 |
Month 2 |
NG 9-1-1 System Component |
Number of Incidents |
Number of Restarts |
Total Down Time |
Number of Incidents |
Number of Restarts |
Total Down Time |
|
Emergency Call Access Technology
|
Legacy Telephony Gateway |
|
|
|
|
|
|
Asterix Telematics Gateway |
|
|
|
|
|
|
Cellular Media Gateway |
|
|
|
|
|
|
SMS Media Gateway |
|
|
|
|
|
|
IP SIP Gateway |
|
|
|
|
|
|
Legacy Wireline SIP Gateway |
|
|
|
|
|
|
Telematics SIP Gateway |
|
|
|
|
|
|
Cellular SIP Gateway |
|
|
|
|
|
|
SMS SIP Gateway |
|
|
|
|
|
|
National LoST Server |
|
|
|
|
|
|
Simulated ALI / MPC DB |
|
|
|
|
|
|
IP LIS (RedSky) |
|
|
|
|
|
|
|
Emergency Call Routing
|
ESRP Server #1 |
|
|
|
|
|
|
ESRP Server #2 |
|
|
|
|
|
|
Local LoST Server |
|
|
|
|
|
|
Business Rules DB Server |
|
|
|
|
|
|
|
Emergency Call Termination Components
|
PSAP #1 IP ACD |
|
|
|
|
|
|
PSAP #2 IP ACD |
|
|
|
|
|
|
PSAP #3 IP ACD |
|
|
|
|
|
|
PSAP #4 IP ACD |
|
|
|
|
|
|
PSAP #5 IP ACD |
|
|
|
|
|
|
|
Network Infrastructure
|
Booz Allen Router |
|
|
|
|
|
|
Texas A&M Router |
|
|
|
|
|
|
Columbia Router |
|
|
|
|
|
|
PSAP #1 Router |
|
|
|
|
|
|
PSAP #2 Router |
|
|
|
|
|
|
|
Network Interfaces
|
Booz Allen Lab |
|
|
|
|
|
|
Texas A&M lab |
|
|
|
|
|
|
Columbia Lab |
|
|
|
|
|
|
PSAP #1 |
|
|
|
|
|
|
PSAP #2 |
|
|
|
|
|
|
PSAP #3 |
|
|
|
|
|
|
PSAP #4 |
|
|
|
|
|
|
PSAP #5 |
|
|
|
|
|
|
|
Overall System Availability
|
System A | |
