Managing capacity in DOCSIS multi-service nets, Part 2
Editor's note: Part one of this article (see Oct.CED, page 24) introduced new ways to monitor multi-service capacity resources using recent enhancements to DOCSIS network management capabilities. Part two of this article will build upon the tools described in part one to explore how capacity monitoring data might be used as input to an overall capacity management model with an MSO's operations support system.
A second key function of multi-service capacity management, capacity planning, provides the ability to estimate the number of services and subscribers that can be accommodated within a specific segment of the DOCSIS access network. A comprehensive capacity planning methodology for multi-service networks is necessary for cable operators to understand the complex relationships between MSO business goals, policy, service definitions, QoS underpinnings, subscriber behavior, and the DOCSIS platform that connects them.
Capacity planning and network modeling is nothing new. Various approaches and models for capacity planning have been proposed for packet-based multi-service environments including both DOCSIS and analogous radio access networks . Though varied in detail and level of complexity, they all share the same common goal: to evaluate what capacity resources will be needed to support subscribers. All methodologies and models are largely common when considering the same basic parameters as inputs and outputs.
DOCSIS capacity planning inputs. The new data collected from the network using the new MIBs and IPDR/SP tools provided by the DOCSIS CMTS furnish many of the inputs required by a given capacity planning methodology. For each segment of the network, the following data is considered by all capacity models and methodologies.
- Service characterization–The behavior of each service type in terms of service throughput and QoS parameters and the associated packet size distribution. For example, VoIP is characterized by relatively short packets at a constant rate. Conversely, a video streaming application may have different packets sizes at variable rates.
- CMTS configuration–Based on DOCSIS CMTS configuration and performance, the maximum bit and packet rate of both upstream and downstream channels will vary. As a result, the actual capacity of the DOCSIS segment is the function of a complex recipe including channel configuration, modulation profile, security configuration, and other DOCSIS overhead.
- Capacity utilization–The current load on the network in terms of devices, services, and reserved bandwidth.
- Service loading and access topology–The total load (by service) definition on the network. Subscriber count by specific service based on devices mapped to topology within the network. For an access network during a specified time interval, such as the busy hour, the percentage of time that service (flow) activity is present. This provides the worst case for estimation of maximum utilization for a specific service as a contributor to overall capacity utilization.
- Service growth projections–For a given market or system, the rate at which each service is projected to grow over a given window of study.
DOCSIS capacity modeling output. The output of any multi-service capacity model describes the estimated utilization of the network as a percentage of overall resource availability to enable congestion prediction, service design and service trending. The capacity planning information rendered by even a simple model can provide the cable operator with valuable insight into the allocation of resources in the access segment.
Figure 1: Example capacity management OSS architecture.
Before an operator adopts an existing capacity modeling and planning methodology, or prefers one of their own, it is important to keep in mind that the goal of the exercise is to provide simple, accurate, timely, and sufficient multi-service planning data.Pulling it all together–Updating the OSS
With these new ways to monitor the DOCSIS CMTS and plan capacity defined, the question becomes–how are these systems integrated into existing operational tools, practices and procedures?
Though each operator has built a different cable operational support system, many core functional elements remain the same. Figure 1 illustrates a generic OSS architecture that includes a Capacity Management function, highlighting the functions described in the form of layers.
Extending the cable OSS monitoring layer will require the addition of the new DOCSIS MIBs to the CMTS SNMP collection system within the platform. Because the SNMP-based collection is from the CMTS and is largely based on the monitoring of interfaces (RF-UPSTREAM, RF-DOWNSTREAM, RF-MAC) the amount of new traffic generated for polling is minimal.
The IPDR/SP subsystem for collection of service usage records from the CMTS is new to most cable OSS infrastructures. As a result, an IPDR/SP collector must be integrated into the monitoring layer. Operators may choose from a number of vendors supporting IPDR/SP or develop their own solutions in-house.
The capacity planning layer is where the greatest amount of flexibility and innovation resides. There exist a number of modeling frameworks, both commercial and open-source , to assist in development. The critical task for operators is to develop interfaces and systems that enable the flow of capacity monitoring data to the capacity planning systems in order to be available to all operational and business support systems requiring it.
This article has provided an introductory survey of the new DOCSIS CMTS platform capacity management features now available for the cable operator to assist in multi-service deployments. In conjunction with a simple capacity management methodology, these tools offer great promise to enhance the cable operator's visibility and predictability of capacity resource requirements and costs.
There exists no single tool providing a "silver bullet" for easily addressing all aspects of multi-service capacity management. Though the DOCSIS platform is a common technology across the industry, each operator will have unique service offerings, goals, engineering and operational standards, and back-office systems–each requiring a unique approach to capacity management.
 J. Chapman, "Multimedia Traffic Engineering...," SCTE Emerging Tech 2002.
 B. Bekele, J. Schnitzer, "Capacity Management in Multimedia Networks," SCTE Emerging Technologies 2005.
 CableLabs, www.cablelabs.com 
 H. Ramzi, "UMTS Radio Access Network Dimensioning," Alcatel Telecommunications Review, 1Q 2001.
 "Validating an 'ns' Simulation Model of the DOCSIS Protocol," Jim Martin, Dept. of Computer Science, Clemson.
 The Network Simulator (NS2), www.isi.edu/nsnam/ns 
 Opnet Corp., www.opnet.com