The main parameters of policy-based auto-scaling are threshold and min/max values. Threshold is the value of a metric that triggers the scaling action, such as CPU utilization or memory usage1. Min/max values are the minimum and maximum number of instances or resources that can be scaled up or down1. Priority and description are not parameters of policy-based auto-scaling. Priority is the order of execution of policies in case of conflicting or overlapping conditions2. Description is the optional text that provides additional information about the policy2. References: 1: Nokia Bell Labs Distributed Cloud Networks, Unit 4: Operating Your Cloud, slide 24 2: Nokia CloudBand Application Manager User Guide, Chapter 9: Scaling Policies, page 151

Microservices are a design principle for cloud-native applications, where each application is composed of a set of small, independent, and loosely coupled services that communicate with each other through well-defined interfaces1. Microservices have the following characteristics2:

• Microservices are chained in order to build an application, meaning that they are connected by a network and exchange messages or data to perform a specific function or task.
• Microservices are differentiated by function, meaning that each microservice has a single responsibility and performs a distinct function within the application.
• Microservices are derived from monolithic applications, meaning that they are created by breaking down a large and complex application into smaller and simpler pieces that can be developed, deployed, and scaled independently.
• Microservices are scalable and resilient, meaning that they can handle variable workloads and recover from failures without affecting the whole application.

Therefore, the correct statements are A and C, while B and D are incorrect. References: 1: Nokia Bell Labs Distributed Cloud Networks, Unit 2: Cloud Technologies and Features, Topic: Microservices and Containerization 2: Module by Module - Self Study Note Guide, DC2.4– Microservices and Containerization

On-premise Edge Cloud and Metro Edge Cloud are the cloud deployments that provide the lowest latency. Latency is the time it takes for data to travel from the source to the destination1. On-premise Edge Cloud is a cloud deployment that is located within the premises of the end-user, such as a factory, a hospital, or a campus2. Metro Edge Cloud is a cloud deployment that is located within the same metropolitan area as the end-user, such as a city or a suburb3. Both On-premise Edge Cloud and Metro Edge Cloud reduce the distance and the number of hops that data has to travel, resulting in lower latency and higher performance4. Far Edge Cloud and Central Cloud are not the cloud deployments that provide the lowest latency. Far Edge Cloud is a cloud deployment that is located at the edge of the operator’s network, such as a regional data center or a base station3. Central Cloud is a cloud deployment that is located at the core of the operator’s network, such as a national data center or a cloud provider3. Both Far Edge Cloud and Central Cloud increase the distance and the number of hops that data has to travel, resulting in higher latency and lower performance4.

References: 1: Nokia Bell Labs Distributed Cloud Networks, Unit 1: Cloud Ecosystem, slide 3 2: Nokia Bell Labs Distributed Cloud Networks, Unit 1: Cloud Ecosystem, slide 14 3: Nokia Bell Labs Distributed Cloud Networks, Unit 1: Cloud Ecosystem, slide 15 4: Nokia Bell Labs Distributed Cloud Networks, Unit 1: Cloud Ecosystem, slide 16

 A logical space in the host’s operating system and B. Is in an isolated environment. A container is a software package that bundles an application and its dependencies, and runs as an isolated process in the host’s operating system1. Containers are lightweight, portable, and scalable, and can be deployed on any cloud infrastructure2. Unlike virtual machines (VMs), containers do not need a separate operating system or hypervisor, and share the kernel of the host3. However, containers have limited access to resources, as they are restricted by the host’s capabilities and the container runtime4. References: 1: Nokia Bell Labs Distributed Cloud Networks, Unit 2: Cloud Technologies and Features, Section 2.2: Cloud Virtualization 2: Nokia Container Services, Overview 3: Nokia Cloud Platform, Overview 4: Containers vs. virtual machines: What’s the difference?, Comparison

Software as a Service (SaaS) is a cloud computing offering that provides users with access to a vendor's cloud-based software. Users do not install applications on their local devices. Instead, the applications reside on a remote cloud network accessed through the web or an API. Within this context, Database-as-a-Service (DBaaS) is considered a form of SaaS, where the service provider manages the database's installation, maintenance, backup, and scaling needs on behalf of the customer. Platform-as-a-Service (PaaS) provides a platform allowing customers to develop, run, and manage applications without dealing with the complexity of building and maintaining the infrastructure typically associated with developing and launching an app. PaaS can include various software components, including database management systems, and thus can be considered a subset of SaaS.

 The 5G Core control plane is responsible for managing the network functions, services, and resources, such as authentication, authorization, session management, policy control, mobility management, etc.1. The 5G Core control plane can be deployed in a distributed cloud architecture, which consists of different cloud layers that have different characteristics and roles2. The Metro Edge Cloud is a cloud layer that is located near the access network, and provides low latency, high bandwidth, and local processing for the control plane functions3. The Core/Central Cloud is a cloud layer that is located in the core network, and provides high availability, scalability, and security for the control plane functions4. The distributed cloud architecture enables the 5G Core control plane to optimize the network performance, efficiency, and flexibility, and to support various use cases and applications. References: 1: Nokia Bell Labs Distributed Cloud Networks, Unit 1: Cloud Ecosystem, Section 1.2: Cloud Architecture 2: Nokia Bell Labs Distributed Cloud Networks, Unit 1: Cloud Ecosystem, Section 1.3: Distributed Cloud 3: Nokia Bell Labs Distributed Cloud Networks, Unit 3: Cloud Resource Planning, Section 3.2: Cloud Resource Planning for Distributed Cloud 4: Nokia Bell Labs Distributed Cloud Networks, Unit 3: Cloud Resource Planning, Section 3.3: Cloud Resource Planning for Central Cloud : Nokia Bell Labs Distributed Cloud Networks, Unit 5: New Services Automation, Section 5.1: Distributed Cloud Services Automation

Habana and EyeQ are two examples of AI hardware solutions that are designed to accelerate deep learning and computer vision applications. Habana is a company that produces AI processors for both training and inference workloads1. EyeQ is a family of system-on-chips (SoCs) developed by Mobileye for advanced driver-assistance systems (ADAS) and autonomous driving2. DataLake and Sparks are not AI hardware solutions. DataLake is a term that refers to a large-scale data storage architecture that can store structured, semi-structured, and unstructured data from various sources3. Sparks is a misspelling of Spark, which is an open-source distributed computing framework for large-scale data processing and machine learning4. References: 1: Nokia Bell Labs Distributed Cloud Networks, Unit 4: AI/ML in Cloud Networks, Section 4.2: AI Hardware 2: Nokia Bell Labs Distributed Cloud Networks, Unit 4: AI/ML in Cloud Networks, Section 4.2: AI Hardware 3: Nokia Bell Labs Distributed Cloud Networks, Unit 2: Cloud Technologies and Features, Section 2.4: Cloud Storage 4: Nokia Bell Labs Distributed Cloud Networks, Unit 4: AI/ML in Cloud Networks, Section 4.3: AI Software

The properties related to a private cloud are high CAPEX and full control. CAPEX stands for capital expenditure, which is the money spent on acquiring or upgrading physical assets such as servers, storage, and network equipment. A private cloud requires high CAPEX because the cloud owner has to invest in building and maintaining the cloud infrastructure. Full control means that the cloud owner has the exclusive authority and responsibility over the cloud resources, security, and operations. A private cloud offers full control because the cloud owner can customize the cloud according to their specific needs and preferences, and can ensure the privacy and compliance of their data and applications. References: Nokia Cloud Platform, Module by Module - Self Study Note Guide