Modification Number: PS11

Contract No. GS00T99NRD2001

Effective Date: April 20, 2000

3.2Frame Relay Service (C.2.3.2)

FRS provides reliable, high speed connectivity between user locations at guaranteed performance levels. The flexibility and reliability of the service make it an attractive alternative to private line networks. The following sections provide the requirements for FRS.

Sprint's Frame Relay Service (FRS) is the flagship data product for commercial and Government customers today. The network is a multiprotocol packet-based data communication network providing high-speed, protocol-transparent switched connectivity. The network architecture is engineered for high performance, flexibility, and reliability.

Service Delivery and Network Architecture and Design(L.38.1.1(a))

(L.38.1.1(a)) The overall network architecture, including the types and capacity of the transmission and switching media, the transmission facility(ies) configuration, and the type of equipment used in its network.

Sprint’s frame relay network architecture is hierarchical consisting of two tiers called the Access layer and the Transport Layer. X

Access Layer

The Access Layer is comprised of high-speed frame relay switches, called Access SwitchesX

XXTransport Layer

The core of the Sprint’s frame relay network, the Transport Layer, is Sprint’s fully meshed OC12 Asynchronous Transfer Mode (ATM) network. The ATM network establishes a reliable high-speed infrastructure for frame relay. This architecture of frame relay switches on the fringes and ATM at the core utilizes Sprint’s Synchronous Optical Network (SONET) transport network. With Sprint’s implementation of SONET technology, the network is able to route around failures almost instantaneously to provide superior survivability and reliability.

Equipment Types

Sprint uses high-speed frame relay switches in the access and transport layers. Access switches provide connectivity up to DS3 frame relay speeds, and interface via ATM to transport layer switches. All switches in the network have full power supply and common logic redundancy as well as one for N redundancy on the access port cards. The switches are located in Sprint's central offices, which are fully protected with both Uninterruptible Power

Supply (UPS) and Emergency Power Supply (EPS). In addition, all switches have built-in shocks to add another level of survivability in case of an earthquake.

Rationale for Network Architecture and Design (L.38.1.1 (b))

Sprint’s hierarchical design modularizes the complex and large frame relay network into component elements of functionality. The key functional elements of Sprint’s network design, as discussed earlier, are access and transport. The advantages of Sprint’s hierarchical network architecture are:

• Scalability

• Manageability

• Optimization of performance.

Scalability

Scalability is the primary advantage of a hierarchical network. Sprint’s hierarchical design is more scalable because it segments the network into smaller components that can easily grow without encountering the difficulties associated with flat or linear architectures. XManageability

The hierarchical network design offers several management advantages. Partitioning the network into smaller elements reduces the complexity of the large frame relay network. X

XXCongestion and Flow Control (L.38.1.1 (c))

(L.38.1.1(c) Congestion and flow control strategy including redundant switch and transmission facilities, control mechanisms, and the degree of flexibility inherent in the architectural design to handle predicted and unpredicted increased traffic loads and/or switch and transmission failures.

Switch and Transmission Redundancy

As described previously, the first line of defense in congestion and flow control is the FRS inherent resilience in the event of switch or transmission facility failure. All switch hardware is fully redundant, so single component failure will not introduce congestion.

Sprint Frame Relay network has several congestion and flow control mechanisms. The primary objective of these mechanisms is to maintain the highest quality of service by minimizing frame discards, warning end-users in advance of congestion, and minimizing the possibility of one end-user consuming network resources at the expense of other users. The aspects of frame relay congestion and flow control mechanisms listed below are described in detail in subsequent sections:

• Explicit Congestion Notification

• Congestion Avoidance Policy

• CIR Rate Enforcement

• Rate Adaptation

Explicit Congestion Notification

Explicit congestion notification (ECN) bits are two bits in the frame header that may be set by the network to notify the user that frames are encountering congestion. The objective of ECN is that users will respond by reducing the flow of data and as such, reduce the risk of

frame discard by the network. There are two ECN bits: FECN (Forward Explicit Congestion Notification) and BECN (Backward Explicit Congestion Notification).

Internally, congestion notification is communicated through the network by setting the bits, called FCI (Forward Congestion Indication) and BCI (Backward Congestion IndicationXXXXXXXXXThe FECN bit is used to warn the user that the frames it is receiving have encountered network congestion. Some application protocols use the destination end to control the amount of traffic the sender is allowed to transmit. In this case, the destination end can instruct the sender to slow down transmission of data in response to receiving FECN bits from the network. The BECN bit is used to warn the sender that the frames it is transmitting are encountering congestion. The sending end may then adjust its flow of data accordingly.

Congestion Avoidance Policy

The discard priority determines which frames should be discarded in preference to other frames during network congestion. Depending on the level of severity of the congestion, the network will choose oneXXXdiscard schemesXXXX.

FECN and BECN indication is only triggered when congestion conditions are encountered in the network, XXXXXXSprint adheres to the industry standards in supporting Committed Information Rate (CIR) which is defined as a minimum sustained data rate with the capacity of bursting up to the full access channel speed. The committed rate measurement interval, abbreviated Tc, specifies the increment of time for which the network continuously measure its information flow. XXThe Excess Information Rate (EIR) is the maximum uncommitted data rate that the network will attempt to deliver during a time internal Tc. XXThe network will attempt to deliver data at the channel access rate (that is equal to CIR + EIR). XXIf network conditions are such that Be>0 and data traffic is injected into the network at a rate greater than CIR, the network will set the DE bit in each frame carrying “excess” data. If a user sends data greater than CIR + EIR, the frames will be discarded at the access point.

Rate enforcement is controlled on a per PVC basis. Each PVC on a frame relay port can be assigned a CIR value which reflects the sustainable throughput the network will support under normal conditions. Throughput in excess of CIR is not allowed, (assuming Be=0).

This means that a single user on a PVC cannot monopolize the available throughput on the port. In this way, users or applications on different PVCs on a single frame relay port have fair access to network resources.XXUnder normal conditions, the user is allowed to send data at the provisioned rates. The frame relay service monitors the subnet for feedback of congestion information. If congestion is detected, the service will XXXXXXXXXXXXXX continue to monitor the subnet for further changes in congestion. Depending on the option, AIR will be either EIR or CIR. Based on this evaluation, it will decide whether the congestion condition has abated or not. If the congestion persists, the service will reduce by 25 percent of the current value of the AIR again. When the congestion condition has abated, the service will be XXXXXXXXXAllowed information rate (AIR) is compatible with existing ITU and ANSI standards.X

XXXXXXXXSprint Equipment at Government Locations (L.38.1.1(d))

Sprint’s FRS may require equipment at the Government’s Service Delivery Point (SDP). As illustrated in Figure 1.B.3-8, FRS service requires a router or FRAD and a physical termination device. Sprint’s FRS may require the following equipment at the Government’s SDP:

• Router

• FRAD

• Power Conversion Equipment

• Physical Termination Devices:

– Channel Bank

– Channel Service Unit/Data Service Unit (CSU/DSU)

– Inverse Multiplexer (IMUX)

– Intelligent Digital Service Unit (IDSU).

Frame Relay Access Device (FRAD)

The Frame Relay Access Device (FRAD) interfaces between non-frame relay protocol devices and the frame relay network. The FRAD also provides the Link Management Interface (LMI) required for frame relay access connections. Figure 1.B.3-8 illustrates the FRAD acting as the interface.

Router

A router is a special-purpose, dedicated device that attaches to two or more local area networks (LANs) and encapsulates multiple protocols into frame relay frames. The networks can support different LAN protocols. The router also provides the necessary LMI signaling required frame relay access connections. The power and environmental requirements for a typical router are detailed in Table 1.B.3-9.

Table 1.B.3-9 Power and Environmental Requirements for a Router
Mount: / Rack Mountable
Power Supply: / -48 Volts dc nominal
Equipment size: / 1.75” x 17.5” x 10.56” nominal
Temperature: / 40 to 85 degrees Celsius
Humidity: / 95 percent (non-condensing)

Frame Relay Access Device (FRAD)

Table 1.B.3-10Frame Relay Access Device (FRAD) Environmental and Space Requirements

Mount: / Stand-alone or Rack Mountable
Power Supply: / -48 Volts dc nominal
Equipment size: / 1.75” x 17.5” x 10.56” nominal
Temperature: / -40 to 85 degrees Celsius
Humidity: / 0to 95 percent (non-condensing)

Power Conversion Equipment

Power distribution may be required in instances where there are large amounts of specialized equipment located at Government locations. Power, in the simplest form, will require a standard 110-120 Vac/15A grounded plug into a wall receptacle. The floor space and power requirements for these power conversion units detailed in Table 1.B.3-11

Table 1.B.3-11Floor Space and Power Requirements

Mount: / Rack-mounted
Equipment size: / 22-1/2” x 16” nominal
Power: / 109-125 Volts AC
Temperature: / 0 to 50 degrees Celsius
Humidity: / 95 percent (non-condensing)

Physical Termination Devices

Physical Termination Devices include Channel Banks, Channel Service Unit/Data Service Units (CSU/DSUs), Inverse Multiplexers (IMUXs), and Intelligent Digital Service Units (IDSUs). Sprint will assist the Government in selecting the appropriate Physical Termination Device to meet the FRS requirements.

Channel Bank (CB)

Channel Banks are employed to provide DS1 circuit termination at a Government location requiring mostly voice service and minimal data services. Sprint provides integrated voice and data service via a single channelized T1 with ESF, as specified by the Bellcore Pub: SR-TSV-002275 and the ANSI T1.102/107/403 standards. Channel Banks multiplex and demultiplex the DS0 sub-channels into a DS1 channelized or fractional DS1 channelized circuit to support voice and data traffic, as illustrated in Figure 1.B.3-9. The typical power and environmental requirements for a typical Channel Bank are listed in Table 1.B.3-12.

Table 1.B.3-12 Power and Environmental Requirements for Channel Bank

Mount: / Rack-mounted or stand-alone
Equipment Size: / 22-3/8” x 15” nominal
Power: / -48 Volts DC nominal
Temperature: / 0 to 50 degrees Celsius
Humidity: / 95 percent (non-condensing)

Channel Service Unit/Data Service Unit (CSU/DSU)

The CSU/DSU is required for service monitoring, troubleshooting and signal conversion and to provide the proper interface for DS1 circuit termination. The CSU/DSU provided will be resident on Government sites. CSU/DSUs are required for data services at fractional DS1, 64/56 kbps, 19.2 kbps, 9.6 kbps, and 4.8 kbps data rates. The typical power and environmental requirements for a typical CSU/DSU are listed in Table 1.B.3-13.

Table 1.3-13Power and Environmental Requirements for CSU/ DSU

Mount: / stand-alone
Equipment size: / 3.125” X 12” X 8.5” nominal
Power: / 115 Volts AC
Temperature: / 0 to 45 degrees Celsius
Humidity: / 95 percent (non-condensing)

Inverse Multiplexer (IMUX)

Inverse Multiplexers are required at the Government’s SDP for access ranging from 6 to 12 Mbps service on Sprint’s FRS. The typical power and environmental requirements for a typical IMUX are in Table 1.B.3-14.

Table 1.B.3-14Power and Environmental Requirements for IMUX

Mount: / Rack-mounted or stand-alone
Equipment size: / 17.2” X 2.8” X 11” nominal
Power: / 120 Volts AC
Temperature: / 0 to 50 degrees Celsius
Humidity: / 0to 95 percent (non-condensing)

Intelligent Digital Service Unit (IDSU)

The Intelligent Digital Service Unit (IDSU) are required at the Government’s SDP for Sprint’s FRS for DS3 (45 Mbps) data rate. The IDSU provides the physical layer interface between the router and the network to support DS3 circuit. The typical power and environmental requirements for a typical IMUX are in Table 1.B.3-15.

Table 1.B.3-15Power and Environment Requirements for IDSU

Mount: / Rack-mounted or stand-alone
Equipment size: / 17.2” X 2.8” X 11” nominal
Power: / 120 Volts AC
Temperature: / 0 to 45 degrees Celsius
Humidity: / 0to 95 percent (non-condensing)

Traffic Performance Calculations and Impact Assessment (L.38.1.1(e))

(L.38.1.1(e)) Traffic calculations that indicate network and service performance during estimated normal, 10 percent, 25 percent, and 50 percent above the estimated normal FTS2001 traffic loads and the means to ensure achieving the required performance as specified in this solicitation.

Sprint’s network can pass the toughest stress test under any predictable GSA loading scenario. Our network will function without effect on quality of performance when various percentages of loading increase are applied to projected FTS2001 average switched data service loading. The increase in traffic calculations indicate that network and service performance will not be affected at usage levels of even up to 50 percent above the estimated normal FTS2001 traffic loads. XXXX

XXXX

Impact of Various Levels of Feature Usage on Service Performance (L.38.1.1(f))

(L.38.1.1(d)) An assessment of the impact of various levels of feature usage on service performance.

Sprint engineers the network to ensure that peak utilization never exceeds percent of network capacity. XXXFrame relay featuresXXare including in determining the bandwidth necessary to ensure that the network maintains a peak utilization XXManaging the network bandwidth and resources in this manner minimizes any impact a feature has on performance of the network.

Network Control and Diagnostics (L.38.1.1(g))

(L.38.1.1(g)) A description of network control and diagnostic capabilities and systems, including equipment and procedures for monitoring and testing each of the services and associated features.

Sprint’s Public Data Service Center (PDSC) in Reston, VA, is the network management hub for all Sprint domestic packet/frame data services and applications. The PDSC contains several Network Control Centers (NCCs) and technical support groups, which configure, monitor, and maintain software and services.

The Service Management Center (SMC) serves as the primary customer contact for any Sprint service issues. When a problem involving a packet data service is reported to the SMC, the SMC opens a trouble ticket and immediately routes it to the PDSC. PDSC technicians are responsible for fault isolation and repair.

The PDSC is committed to providing the best possible service for packet data services. Through the creative use of leading edge technology, Sprint has deployed new systems that offer benefits including a great increase in the PDSC’s metabolism—the speed at which we can launch new products and the power to absorb new technologies.

These new systems and processes allow Sprint to provide:

• Dramatic increases in our ability to pro-actively detect, manage, and, in many cases, automatically resolve problems

• Automatic testing and evaluation of every problem

• Automation of routine problem resolution

• Significant decreases in overall problem resolution time

• Enhancement of Sprint’s ability to manage trouble tickets

• Fast analysis and correlation of historical information for network capacity planning and traffic management

• Easy customization of support offerings, based upon user’s needs

• Automatic enforcement of optimized ticket management policies for the fastest possible problem resolution.

XXX

XXXX

Network Transmission and Synchronization Plans (L.38.1.1(h))

(L.38.1.1(h)) Network transmission and synchronization plans for the various services.

Network synchronization, the communication timing within and between networks, plays a critical role as high-speed communication networks span the globe. Sprint is particularly well suited to handle the Government’s demand for worldwide data transmission because of our 100 percent digital, fiber-optic network in the U.S.—the ideal transmission medium for high-speed data transmission.

Digital network synchronization as implemented on the Sprint network, assures that the network can transport data end-to-end and interconnect with other networks with minimal degradation. The benefits of Sprint synchronization include transmission clarity, virtually error-free operation, and survivability using Digital Cross-connect System (DCS) devices.

Sprint uses either a LORAN (long range navigation) or a GPS (Global Positioning Satellite) Primary Reference Source at all of our nodes and switch sites to the Government with state of the art plesiochronous synchronization. XXAs the first U.S. carrier to provide 100 percent fiber-optic transport with plesiochronous synchronization, Sprint delivers advanced, reliable telecommunications services. The Sprint network uses a plesiochronous method of providing Stratum 1 timing sources at different network nodes rather than one centralized source. The plesiochronous technique does not experience timing degradation of long timing distribution links. This synchronization method is uniquely suitable to a total fiber-optic network because of the automatic reconfiguration capability provided by diverse fiber links from a node to different Stratum 1 clocks, which are provided via Loran-C systems.