Scaling Management

Ericsson Service-Aware Policy Controller

1	Introduction
1.1	Document Purpose and Scope
1.2	Revision Information
2	Functions and Concepts
2.1	Types of Operation
3	Managed Object Model

1 Introduction

1.1 Document Purpose and Scope

This document provides an overview of the management model and concepts associated with the Scaling Management managed area.

A managed area is represented by a group of Managed Object Classes (MOCs) within the Managed Object Model (MOM).

1.2 Revision Information

Rev. A	This is the first release of this document.

2 Functions and Concepts

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

Scale-out, that is, adding one or more Virtual Machines (VMs) when a cluster requires more processing.
Graceful scale-in, that is, removing one or more VMs nodes when a cluster needs less processing. Released resources can be allocated again when more processing is required.

A scale-out operation is performed by adding one or more VMs to the Virtualized Network Function (VNF) cluster. Preboot Execution Environment (PXE) boot is used to distribute software to new VMs. 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. Performance counters can be used as input to decide which scaling operation is to be performed.

A VM can have one of the following two roles:

System
When a VM works in role System, it cannot be scaled in or scaled out, as it has a system-defined size and a system-defined role. An example is that two VMs are allocated to the Operation and Maintenance (O&M) role within the VNF, which is named SC-1 and SC-2. These VMs cannot be scaled in or scaled out.
Default-Role
When a VM works in Default-Role, it does not have a specialized role like, for example, SC-1 or SC-2. This means that VMs with Default-Role can be scaled in or scaled out, depending on the capacity needs of the VNF.

2.1 Types of Operation

Scaling Management supports the following operations:

Scale-out
The scale-out operation adds a node to the cluster. Further details on how to perform this operation depend on the deployment type:
- Deployment in a Cloud Data Center: See the procedure in Configure Scale-Out.
- Deployment in specific HW: See the procedure in SAPC Expansion Instruction.
Graceful scale-in
The graceful scale-in operation, intended for Cloud scenarios, removes a node from the cluster. Further details on how to perform this operation can be found in Configure Graceful Scale-In.

3 Managed Object Model

The Scaling Management managed area is represented in the Managed Object Model (MOM) as follows:

ManagedElement
   +-SystemFunction
      +-SysM
         +-CrM
            +-ComputeResourceRole
            +-Role

For general information about the MOM, MOCs, Managed Objects (MOs), cardinality, and related concepts, refer to Managed Object Model User Guide.

The Scaling Management MOCs are described in Table 1.

Table 1 Scaling Management Managed Object Class Descriptions
Managed Object Class	Description
CrM	The root of the Scaling Management model.
ComputeResourceRole	Describes the service (defined by the role) provided by the compute resource.
Role	Describes a role that determines the type of service provided by a compute resource.