CSCF Scaling Management
Call Session Control Function

Contents

1Understanding Scaling Management
1.1Key Scaling Management Concepts

2
Basic Scaling Management Operations

1   Understanding Scaling Management

Scaling Management provides a management interface to configure the following on the Managed Element (ME):

The Scaling Management managed area can be found under the CrM Managed Object Class (MOC) in the Managed Object Model (MOM). For general information about the MOM, MOCs, cardinality, and related concepts, see Managed Object Model User Guide.

When the system starts a scaling operation, it enters in the Maintenance Mode, meaning the overload regulation is lowered to the vDicos initial configuration parameter LOAD_REG_MAINT_LIMIT.

The scaling operation involves planned reconfiguration of distribution units. This activity is performed in the quickest possible manner with high priority. Hence, load regulation-related alarms can appear during scaling operation. Such alarms are not expected to be present for longer than 2–4 seconds. The effect is minimal on traffic handling capability.

A scale-out operation is performed by adding one or more VMs to the Virtual Network Function (VNF) cluster, see Section 1.1.1 Auto Scale-Out. For graceful scale-in, the VNF cluster reallocates the resources from VMs to be scaled-in and moves to other VMs to prevent data loss, see Section 1.1.2 Graceful Scale-In. Performance counters can be used as input to decide which scaling operation is to be performed.

A VM is assigned a role with an attribute that describes the scaling behavior:

1.1   Key Scaling Management Concepts

1.1.1   Auto Scale-Out

Auto Scale-Out is an operation where one or more new compute resources are launched, see Figure 1. The system automatically detects, configures, and brings up the nodes as a member of the scaling domain of the cluster. See Figure 2 for an example where one new compute node is added to the cluster.

Figure 1   A New Compute Resource Is Spawned and Available

Figure 2   After Auto Scale-Out, a New Resource Is Added to the Cluster

1.1.2   Graceful Scale-In

Graceful Scale-In is an operation where one or more compute resources, part of the scaling domain of the cluster (see Figure 3) are removed from the cluster (see Figure 4) to free up resources.

Figure 3   The Node Named PL-(N-1) Is Part of the Cluster

Figure 4   The Node Named PL-(N-1) Is Removed from the Cluster and Its Resources Can Be Released

Note:  
The Graceful Scale-In operation can be rejected by the cluster if, according to the automatic estimation of the system, the target size of the cluster does not have the memory resources to serve the needed memory capabilities for the ongoing traffic.

1.1.3   Forceful Scale-In

Forceful Scale-In is, similarly to Graceful Scale-In, an operation to remove one or more nodes from the scaling domain of the cluster. The only difference is that in this case, either the node is not available (see Figure 5) or scale-in with potential traffic loss is acceptable. If the node is not available, it can be either because it already freed up its resources or because of a failure. Therefore, the removal is only an administrative operation, see Figure 6.

Figure 5   The Node Named PL-(N-1) in the Cluster Scaling Domain Is Unavailable

Figure 6   The Node Named PL-(N-1) Is Removed Administratively from the Cluster

2   Basic Scaling Management Operations

Scaling Management is accessed using NETCONF or the Ericsson Command-Line Interface (ECLI) to manipulate the Management Information Base (MIB). The following operations can be performed:

Legal | © Ericsson AB 2018