MSPP Network Design Methodology


Network designers must address three important questions during the transport network planning process:

What level of reliability must the transport system provide to meet end-user availability requirements?

How will the network be synchronized to ensure that all transmissions from each of its elements are precisely timed?

How will the network be managed to facilitate fault management, provisioning, and performance evaluation of the system?

This section covers these design parameters and the factors to consider when making decisions related to these.

Protection Design

A key feature of MSPPs is the capability to provide a very high level of system reliability because of the multiple forms of redundancy or protection provisionable in the platform. These protection mechanisms are available to the MSPP network designer:

  • Redundant power feeds

  • "Common control" redundancy

  • "Tributary" interface protection

  • Synchronization source redundancy

  • Cable route diversity

  • Multiple shelves, or chassis

  • Protected network topologies (rings)

Redundant Power Feeds

MSPPs require a power source with a -48 V potential (within a certain tolerance) to operate. Two separate connections for power supply cabling are generally provided on the rear of an MSPP chassis. These are typically referred to as Battery A and Battery B. The internal power distribution is designed so that the system can continue to function normally if one of the supply connections is removed or fails. If only one of these supply connections is cabled to the power source, which is typically an alternating current (AC)to direct current (DC) rectifier plant, an alarm is raised in the system's software (such as "Battery Failure B").

For maximum protection from service interruption from power issues, the MSPP design should include connections for both the A and B terminals. Power feed cable sizing and the associated fusing should follow the local electrical codes. If possible, separate power supplies should provide these redundant feeds, to avoid losing both feeds in a single-supply failure. Finally, it is strongly recommended that you use power supplies that include battery back-up systems, to maintain service if a commercial power outage occurs.

Common Control Redundancy

Most fully featured MSPP systems employ a modular architecture. This means that various cards (also called blades or plug-ins) can be inserted into the chassis to provide various functionality or service interfaces. Typically, a subset of the cards installed in the chassis is known as common control cards. These cards provide functions that all the installed interface cards need, and also control and monitor the operation of the system. Consider some examples of the functions that the common control cards in an MSPP provide:

  • System initialization

  • Configuration control

  • Alarm reporting and maintenance

  • System and network communications

  • Synchronization

  • Diagnostic testing

  • Power monitoring

  • Circuit cross-connect setup and maintenance

Because many of these functions are critical to the proper operation of the system, the cards that perform these functions typically must be installed in pairs. For example, the Cisco ONS 15454 MSPP reserves five chassis slots for common control cards; four of these are used to house two redundant pairs. The Timing, Communications, and Control (TCC) card and the Cross-connect (XC) card are essential to the system and are always placed in an active/standby pair. If the active card fails, the standby card takes over. Failure to install the secondary common control card(s) normally results in an alarm condition in the MSPP, such as a "Protection Unit Not Available" alarm.

Tributary Interface Protection

Interface cards that connect an MSPP to external equipment, such as routers, Ethernet switches, and private branch exchange (PBX) switches, are sometimes referred to as tributary interfaces. MSPP cards can be protected to ensure service continuity in case of card failure. MSPPs provide protection in many ways, which include one or more of the following:

  • 1:1 protection A single protect, or standby, interface card protects a single working, or active, service interface card. This type of protection can be provided for electrical time-division multiplexing (TDM) interfaces, such as DS1, DS3, or EC-1. If the working card fails or is removed, traffic switches from the working card to the protection card. The standby card slot is linked through the MSPP chassis backplane to the working card slot so that it can reuse the cabling (such as coax cables for a DS3 interface) from the working slot if a protection switch occurs.

  • 1:N protection In this protection scheme, a single standby card is used to protect multiple working cards (such as a single DS3 interface card used to protect up to five working DS3 cards). This is an advantage economically because fewer total cards are required to protect the working service interfaces. In addition, non-revenue-generating protection cards consume less valuable chassis real estate. As in 1:1 protection, some MSPP systems provide this type of redundancy for electrical TDM interfaces.

  • 1+1 protection Optical (OC-N) interfaces can use this protection type in a failover scenario. 1+1 protection implies that a standby card or port protects a single working optical card (or port on a multiport card). Each of the optical cards must be cabled to the external equipment to provide this functionality.

  • Higher-layer protection Data interfaces, such as native Ethernet cards, are typically unprotected. However, if required, you can protect these interfaces by installing separate cards, with each having a 100-Mbps or Gigabit Ethernet (GigE) connection to a switch or router. This external networking equipment can then be configured to provide protection at Layer 2 or Layer 3. As an example, separate GigE links with load balancing can be configured between the MSPP Ethernet interface and the external switch or router.

  • 0:1 protection This is also known as unprotected operation. This can be an option with certain designs, as in the case of Ethernet or Storage-Area Network (SAN) extension services, and when Layer 3 is used to protect data traffic.

Figure 5-1 gives an example of an MSPP (a Cisco ONS 15454) with multiple service interfaces configured with the various tributary protection methods. For critical service requirements that cannot be subjected to outage caused by the failure of a single interface card, it is strongly recommended that you provide tributary interface protection. In addition to avoiding traffic loss, these protection mechanisms can be used to defer time-consuming repair visits (truck rolls) for card replacements until regularly scheduled maintenance visits.

Figure 5-1. MSPP with Multiple Tributary Protection Methods


Synchronization Source Redundancy

Timing is an essential element of Synchronous Optical Network (SONET)/Synchronous Digital Hierarchy (SDH) networks. MSPP systems enable network operators to provision multiple synchronization sources to provide redundancy. Primary timing can be provided for a node either externally, through wired connections to a digital clocking source, or from a connected OC-N interface. In either case, MSPP systems also must be equipped with internal clocking sources (usually contained within the common control cards) for failover in case the primary sources fail.

For MSPP locations that have a clock source installed, such as a Cesium clock or Global Positioning System (GPS) clock receiver, the MSPP typically provides redundant clock source connection terminals. These are normally referred to as Building Integrated Timing Supply (BITS) connections. These should be wired to diverse timing interfaces on the BITS clock system for maximum redundancy. In this type of arrangement, the network operator can configure three timing sources for the MSPP node and rank them in order of preference to use. For example, the first-choice timing reference source would be the BITS-1 connection. The second choice would be the BITS-2 connection (in case the BITS-1 connection fails), followed by the internal system clock as a last resort.

Similarly, MSPP locations that do not have an external clocking source can be configured with three timing source options. The first two options would be the East and West OC-N ring interfaces, which can be traced back to another node's clocking source. The third option is the internal clock.

Cable Route Diversity

Physical cable route diversity should be considered to ensure service continuity in case of a cable break or inadvertent disconnection. Cable route diversity can be provided on both a system-level and a network-level basis.

At the system or node level, diverse pathing and ducting for direct system connections can be provided during shelf installation, with one cable lead routed away from the shelf on the left side of the network bay frame or cabinet, and the other cable lead on the right. Connections that can provide diversity in this manner include power feeders, timing leads, optical fibers, and Category 5 Ethernet cables.

At the network level, optical cable diversity can be provided for MSPP nodes configured in ring topologies. This protection can be provided both inside the building in which the MSPP is located and in the exterior (or outside-plant) cabling that is used to connect to other node locations. For example, a carrier operating an MSPP network might choose to use diverse routing for each pair of fibers connecting MSPP nodes in a two-fiber ring, such as a unidirectional path switched ring (UPSR). This could include separate underground conduit runs leaving a building and diverse geographical cable routes. If only one route exists between two locations, such as to a "spur" site on a ring, diversity can be provided by using two fibers in a buried fiber cable and two fibers in a cable lashed to a utility pole line. This would be the case when a particular site is limited to a single physical-access route (such as one roadway) from the remaining sites on the network.

Multiple Shelves

In an MSPP network that is used to transport critical services, such as medical or E911 applications, consideration should be given to providing chassis-level protection for service-termination nodes. For example, a hospital using a resilient SONET ring to carry real-time video for remote robotic-assisted surgery might elect to install MSPP nodes for interface diversity in separate data rooms or closets in a building or campus, to protect against a power outage, fire, or flooding hazard.

Protected Network Topologies (Rings)

Protected network topologies, or rings, can be used to protect a network if a single node or fiber span fails. Rings allow protected traffic to be rerouted automatically over an alternate path if the active path becomes unavailable. Ring topologies are covered in detail in the sections "UPSR Networks" and "BLSR Networks."

Network Timing Design

SONET MSPPs rely on highly accurate clocking sources to maintain proper network synchronization. Each element in the network should be timed from a Stratum Level 3 (or better) source traceable to a primary reference source (PRS). The PRS is typically a Stratum Level 1 clock.

Timing Sources

For timing design, the first consideration should be the type of timing source available at each location. As discussed previously, the possible timing sources for MSPPs include the following:

  • External sources, such as a Cesium or GPS clocks

  • Line sources, which include the SONET optical ports

  • Internal clock, which is normally a Stratum 3 clock built into a system common control card

Based on these different timing source types, a network designer has several options to choose from when deciding how the network will be synchronized. The following are the most common recommended designs:

  • External/line-timed configuration In this configuration, one MSPP in the network is timed from an external clocking source, such as a BITS clock; the other MSPP nodes are line-timed from their optical interfaces. The clocking source for the externally timed node is traceable to the PRS. This is a typical timing design for a carrier access ring, in which one (or more than one) node is located in a telco central office and the other nodes are located at remote locations. Figure 5-2 shows this configuration.

    Figure 5-2. External/Line-Timed Network Synchronization Configuration

  • Externally timed configuration If external clocking sources are available at all MSPP locations, each node can be cabled and configured as externally timed. Figure 5-3 shows an example of this type of design. Because this requires clocks at each site, this type of configuration is normally seen in interoffice transport applications, in which each MSPP is located in a telco central office. If all clock sources are traceable to a single primary reference source, the network is said to be synchronous. However, if the various local clocks are traceable to two or more primary reference sources with nearly the same timing (such as an MSPP network that spans more than one carrier), the network is referred to as plesiochronous.

    Figure 5-3. Externally Timed Network Synchronization Configuration

  • Internal/line-timed configuration For networks in which no external clocking source is available, such as a private corporate network, a single MSPP node can be configured for internal timing from its embedded Stratum 3 clock, with the remaining nodes deriving their timing from their connected optical interfaces. Some MSPP vendors refer to configured Internal clocking as "free-running" mode. This type of design is unacceptable for rings in which SONET connections (such as OC-3 or STS-1) are required to other networks. Also, additional jitter is introduced on low-speed add/drop interfaces, such as DS1s. This can cause problems if these interfaces are connected to external equipment that is sensitive to such jitter. One example is a voice switch that contains a Stratum 3 clock of its own. Figure 5-4 depicts a ring configured for the internal/line-timed configuration.

Figure 5-4. Internal/Line-Timed Network Synchronization Configuration


Timing Reference Selection

After selecting the timing configuration, the network designer must assign the prioritized list of timing references to be provisioned for each MSPP node. The MSPP system software uses this prioritized list to determine its primary reference source for timing and the order of failover switching in case the primary source becomes unusable or unavailable.

Typically, this list can be set up in the MSPP node to be revertive or nonrevertive. Revertive means that, if the primary reference fails and a switch to the secondary reference occurs, the node switches back to the primary source after the problem that forced the switch is corrected. A countdown timer is implemented for this reversion switch so that the network operator can set an amount of time (such as 5 minutes) to allow the primary reference to stabilize before the node switches back. This prevents multiple clock source switches from a "flapping" primary timing source.

The order of prioritization for timing sources varies based on whether the node is externally timed, line-timed, or internally (free-running) timed. For externally timed nodes, in which the MSPP is synchronized to DS1 timing leads cabled to the chassis from an external clocking source, the primary and secondary options for timing should be the redundant pair of connections, usually referred to as BITS-1 and BITS-2. Generally, it is recommended that you use the internal Stratum 3 clock as the third option so that if both BITS inputs fail, the MSPP fails over to internal timing. However, some MSPP systems allow for another option, which is to provision an optical port for the third selection. Cisco calls this mixed timing mode. This is an option for a network that is configured in the external/line-timed configuration, in which more than one node uses external timing. Using mixed timing can be tricky, however, as you will see in the examples later in this chapter. Although the Cisco ONS 15454 allows for mixed timing configuration, Cisco does not recommend its use and urges caution if you are implementing it.

For line-timed nodes, a typical recommended prioritization list includes the "main ring" optical ports, East and West, for the first and second options, with the internal clock as the third option. Internally timed nodes should be provisioned to use the internal clock for all three references.

Synchronization Status Messaging

To maintain adequate timing in an MSPP network, SONET provides a protocol known as Synchronization Status Messaging (SSM) to allow for communication related to the quality of timing between nodes. This information is carried over 4 bits in the S1 byte of the line overhead, with each node transmitting its current status to the adjacent line-terminating equipment (LTE). Recall that an LTE is a piece of SONET equipment that can read and modify the line overhead bytes. These messages enable each MSPP in the network to select alternate timing sources, as defined in the prioritized reference list, when events in the network make a change necessary. Two sets of SSM codes are in use today: The older and more widely used version is Generation 1; Generation 2 is a newer version that defines additional quality levels. Table 5-1 defines the message set for Generation 2.

Table 5-1. Synchronization Status Messaging Generation 2 Message Set

Message Description

Message Acronym

S1 Bits 58

Quality Level

Stratum 1 Traceable

PRS

0001

1

SynchronizedTraceability unknown

STU

0000

2

Stratum 2 Traceable

ST2

0111

3

Transit Node Clock

TNC

0100

4

Stratum 3E Traceable

ST3E

1101

5

Stratum 3 Traceable

ST3

1010

6

SONET Minimum Clock Traceable

SMC

1100

7

Stratum 4 Traceable

ST4

8

Do Not Use for Synchronization

DUS

1111

9

Provisionable by Network Operator

PNO

1110

User assignable


To understand how these SSM messages are used in a SONET MSPP network, consider the example network shown in Figure 5-5. This network is configured in the externally timed/line-timed arrangement, with Node 1 being externally timed from a collocated BITS clock and Nodes 26 being line-timed traceable back to the timing source at Node 1. Because Node 1 is being timed by a Stratum 1 clock, the last 4 bits of the S1 byte transmitted in both the East and West directions are set to the value 0001, indicating the primary reference source.

Figure 5-5. SSM Operation


Note

Node 1 has the two BITS inputs as its primary and secondary timing sources, with the internal clock as the last resort.


Proceeding clockwise in the network, Nodes 2, 3, 4, 5, and 6 have been provisioned to select the timing from the OC-N card installed in their West slot (in Figure 5-5, this is Slot 6) as their first choice for timing.

Note that each of these nodes receives the SSM value of PRS on their Slot 6, indicating that the timing being received is of Stratum 1 quality. Each node passes that along in the daisy chain to the next node in the ring.

In the reverse direction (counterclockwise), each MSPP node transmits the SSM value of DUS (Don't Use for Synchronization) back in the direction from which primary timing is received. This is required to prevent a timing loop from occurring somewhere in the network if the primary timing source fails. Note that at the Node 6toNode 1 connection, the PRS value is being both transmitted and received by each of the two network elements. This is a normal scenario. Node 1 always ignores the PRS being received on its Slot 6 OC-N card because that optical line is not among its possible timing references. Meanwhile, Node 6 makes use of the SSM information received on its Slot 12 only if its primary reference source fails.

The example shown in Figure 5-5, in which each of the line-timed nodes has the same Reference 1, is one typical method of timing configuration in externally timed/line-timed designs. Another method for this type of network is to provision the ring OC-N interface that is closest to the primary reference source as Reference 1. For example, in Figure 5-5, Nodes 2 and 3 would have the OC-N interface in Slot 6 as their Reference 1, whereas Nodes 5 and 6 would have Slot 12 for their first selection. Node 4 could be provisioned either way, since it is the same number of "hops" in either direction back to the BITS-connected Node 1. Cisco recommends the latter selection criteria for their MSPP, the ONS 15454.

Network Management Considerations

Network management is an important consideration in MSPP network design. The designer has various options, depending upon the type of network environment in which the equipment is deployed. Fault detection, configuration and provisioning, performance monitoring, and security management are all functions of the network-management system.

In an enterprise campus or privately owned metro MSPP network, each MSPP node can have the capability to connect to the local LAN for remote management. Figure 5-6 shows this type of scenario. From the standpoint of fault recovery, this is an advantageous arrangement because no single network connection failure will isolate the network from the network-management system. Each MSPP node has its own connectivity for management purposes.

Figure 5-6. An MSPP Network with LAN Connections to Each Node for Network Management


Figure 5-7 shows a more typical service provider scenario. Service provider MSPP networks typically have one or more network elements located in secure, company-owned locations, such as a telco central office. The other nodes are normally located on the premises of the service provider's customers or in common locations serving multiple customers. The node(s) located in the service provider's central office (sometimes referred to as a point of presence, or POP) can be linked back to the network operations center (NOC). This node is known as the gateway network element (GNE). The other nodes, called external network elements (ENEs), use the SONET in-band management channel, known as the data communications channel (DCC), to communicate to the NOC via the GNE. In this example, a single MSPP node (Node 1, the GNE) is attached to the management network; the other nodes, ENEs, use the DCC for management connectivity. If additional MSPP nodes in the network are located in facilities that the service provider owns and can also be used as GNEs, the single point of failure for network-management connectivity shown in this example could be eliminated.

Figure 5-7. An MSPP Network with a LAN Connection to a Single GNE Node for Network Management





Building Multiservice Transport Networks
Building Multiservice Transport Networks
ISBN: 1587052202
EAN: 2147483647
Year: 2004
Pages: 140

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net