Untitled Quiz

Cloud Computing Introduction

• Cloud computing is an information technology paradigm enabling ubiquitous access to shared configurable resources.
• It rapidly provisions resources with minimal management, often via the internet.
• Cloud computing relies on resource sharing to achieve economies of scale, similar to a public utility.
• Cloud computing delivers computing services (servers, storage, databases, networking, software, analytics) over the internet.
• Cloud providers bill users based on usage, like gas or electricity.

Course Logistics

• Required textbook: "Cloud Computing, Theory and Practice" by Dan C. Marinescu, published by Morgan Kaufmann.

Course Contents

• Introduction to Cloud Computing
• Virtualization I
• Virtualization II
• MapReduce Batch Processing
• MapReduce in Heterogeneous Environments
• Large-Scale Resource Management
• Data Center Networking
• Cloud Distributed Storage
• Real-Time Data Stream Processing

Lecture Contents

• Defining Cloud Computing
• Early Models of Cloud Computing
• Delivery Models and Services
• Ethical Issues in Cloud Computing
• Cloud Vulnerabilities
• Parallel Computing
• Distributed Systems

What is Cloud Computing?

• Cloud computing is an information technology (IT) paradigm enabling ubiquitous access to shared pools of configurable system resources.
• The resources (servers, storage, databases, networking, software, analytics) are provided over the Internet.
• This provision is characterized by rapid provisioning and minimal management effort.
• This allows for economies of scale, analogous to a public utility.

Cloud Computing Models, Resources, Attributes

• Delivery models: Software as a Service (SaaS), Platform as a Service (PaaS), and Infrastructure as a Service (IaaS).
• Deployment models: Public cloud, Private cloud, Community cloud, and Hybrid cloud.
• Resources: Compute servers, storage servers, networking, and applications.
• Defining attributes: Massive infrastructure, utility computing, pay-per-usage, internet accessibility, and elasticity.

Early Models of Cloud Computing

• Grid computing: Developed by the National Labs in the early-1990s, primarily for scientific computing. A distributed system for large-scale, often non-interactive tasks, involving many files.
• Utility computing: Emerged in 2005-2006, targeting enterprise computing; a service provisioning model where computing resources and infrastructure are made available on-demand, with usage-based billing.

Cloud Computing - Characteristics

• Cloud computing offers scalable and elastic computing and storage services.
• Resources are shared.
• Cloud systems use internet technology.
• "Elastic computing" refers to dynamically acquiring resources on demand, supporting fluctuating workloads.
• Resources are metered, and users are charged accordingly.
• Cloud computing is cost-effective due to resource multiplexing. Cost savings are often passed to the user.

Cloud Computing (cont'd)

• Data storage: Data is stored in the cloud, potentially closer to the location of usage and presented as location-independent to users. This storage strategy enhances reliability, security and reduces communication costs..
• Management: Maintenance and security are handled by service providers. The specialization and centralization involved in this management can optimize efficiency.

Cloud Computing Advantages

• Shared resources: CPU cycles, storage, and network bandwidth are shared. This allows for higher resource utilization from multiple applications without conflicts.
• Multiplexing: Peak demands of multiple applications are not synchronized, thereby maximizing resource utilization.
• Aggregated resources: Resources are aggregated to support data-intensive applications.
• Data sharing: Data sharing enables collaborations among groups across the globe.

Cloud Computing Advantages (cont'd)

• Eliminates initial investment: Eliminates initial investment in private computing infrastructures and reduces maintenance/operational costs.
• Cost reduction: Pay-as-you-go for computing resources reduces costs significantly.
• Elasticity: Adapts to fluctuating workloads with large peak-to-average ratios.
• User convenience: Virtualization creates familiar environments instead of idiosyncratic ones.

Types of Clouds

• Public cloud: Resources are made accessible to the general public or a large industry group and owned by the provider organization.
• Private cloud: Resources are operated solely for a specific organization, often within their own data center.
• Hybrid cloud: Combines public and private clouds, offering a unique entity bound by standardized technology, enabling data and application portability.
• Community cloud: Shared infrastructure shared by multiple organizations that have shared concerns.

Why Cloud Computing is Successful

• Exploits advances: Leverages recent advances in software, networking, storage, and processor technologies. The same companies making these advancements also often provide cloud services.
• Economic reasons: Significant use in enterprise computing, by industry organizations, financial institutions, and government, driving significant economic impact.
• Infrastructure Management: Homogeneous infrastructure (often migrating toward heterogeneous) simplifying security, resource management, and fault tolerance compared to managing heterogeneous resources across multiple entities.

Challenges for Cloud Computing

• Availability: Service provider failures pose risks to service provision.
• Data confidentiality & auditability: Maintaining the security and privacy of user data is a significant hurdle.
• Service diversity & user mobility: Diverse services offered by providers hinder user mobility, causing customers to become reliant upon vendors.
• Data transfer bottlenecks: Applications requiring massive data transfers, or real-time exchanges will face significant challenges.

More Challenges

• Performance unpredictability: A consequence of resource sharing, challenges efficient resource management and reliable QoS (quality of service). How to use resource virtualization, and performance isolation are critical considerations.
• Resource management: Managing multiple workloads running on shared resources is a challenge. Are self-organisation and management solutions feasible in large datacenters?
• Security and confidentiality: Security and confidentiality are critical for sensitive applications, such as healthcare.

Cloud Delivery Models

• SaaS (Software as a Service): High level. Applications are supplied by the service provider. Users do not manage the underlying infrastructure. Email (Gmail), CRM (Salesforce)
• PaaS (Platform as a Service): Applications are deployed using vendor-supported languages. Users manage applications, while the provider manages the infrastructure. Examples: Google App Engine, Windows Azure.
• IaaS (Infrastructure as a Service): Low level. Provides compute resources (CPU, VMs, storage) as needed. Users manage the operating system and applications; the infrastructure provider handles the rest. Example: Amazon EC2.

Cloud Activities

• Service management and provisioning: Including virtualization, service provisioning, call centers, operations, system, QoS, billing and accounting, SLA management, technical support, and backups.
• Security management: Including authentication, certification, intrusion prevention/detection, virus protection, cryptography, physical security, incident response, access control, and firewalls.
• Customer services: including customer assistance, subscriptions, business intelligence, reporting, and customer preference management.
• Integration services: Including data management and development.

Ethical Issues

• Shift in control: Releasing control to third-party services introduces new ethical considerations.
• Data storage locations: Data storage in multiple sites managed by different organizations presents challenges to responsibility and control.
• Interoperability: Multiple services operating across the network raise implications on security threats like unauthorized access and data corruption, as well as infrastructure failure.

De-perimeterisation

• Systems across multiple organizations can pose security risks.
• Cloud complexity makes determining responsibility difficult in case of undesirable events.
• Identity theft and fraud become possible via unauthorized access and new dissemination methods in social media.

Privacy Issues

• Cloud providers store very large quantities of sensitive personal data.
• Privacy acceptance is linked to the privacy policies and practices of the companies and the countries where data centers are located.
• Privacy attitudes vary by culture.

Cloud Vulnerabilities

• Malicious attacks and infrastructure failures, e.g., power outages, can affect clouds.
• Such incidents can disrupt the internet domain name servers, impacting cloud access. Examples: Akamai attack in 2004. Google and Gmail outages in 2009. Amazon outages in 2012.

Parallel Computing

• Parallel computing: Computations are done concurrently. Large problems are broken down into smaller ones for simultaneous solution. It's used to run demanding problems, without the need for a dedicated large system.
• Parallel computing enables processing power that might not otherwise be available on a single system.
• Parallel computing reduces processing time.

Parallel Computing – Amdahl's Law

• Amdahl's Law addresses the theoretical upper limit on the speedup achievable by parallelization of a given program.
• The formula accounts for the fraction of a program's running time which is difficult or impossible to parallelize. Significant value is lost in programs where the parallelization effort is low.

Distributed Systems

• A collection of autonomous computers, networked together for a common goal.
• Computers work together, using middleware to coordinate activities and share resources.
• Characteristics include: perceived unity, autonomous components, each system independently manages scheduling and security. Multiple control points and failure possibilities. Access to resources may not be universal. Scalability via additional resources. Designed for high availability under adverse conditions.

Summary

• Key concepts of cloud computing are covered, including various models, advantages, challenges, ethical considerations, and underlying technologies.