Virtualization Concepts and Hypervisor Functions

Cloud Computing (Part I)

• The presentation was given by Rodrigo Bruno
• Covered the topic of Cloud Computing, part I
• Also touched on the concept of Virtual Machines

Recap (Protocols)

• Presented a diagram illustrating the layered network protocols
• FIRE (7, "Lucifer", evade) protocols appear at the top and bottom layers

Recap (Routing)

• Depicts a network diagram with computers, networks, gateways, switches, and a router
• Computers have unique IP addresses (e.g., 192.168.1.7, 10.80.7.8)
• Network 1 and Network 2 are connected by a router
• Network gateways have IP addresses (e.g., 192.168.1.1, 10.80.7.1)

Recap (OSI Layers)

• Explains the functionalities and responsibilities of each OSI layer
• TCP (Reliable communication) and IP (Multi-network communication) are mentioned
• Application, Presentation, Session, Transport, Network, Data Link, and Physical layers are described

Recap (Internet)

• Showcases how data flow through routers via hops
• Individual routers have unique IP addresses (e.g., 171.64.10.1, 173.255.219.1)
• Diagrams demonstrates different routers on an internet path, including Nick's machine with IP addres (171.64.64.166) and a destination on codingbat.com with IP address (173.255.219.70)

Possible Code Contents

• Presentation included a cartoon-style diagram illustrating common coding problems
• Problems like TODO's, hidden bugs, duplicate code, complexity, missing features, etc. were highlighted

Mainframe Virtualization (early days)

• Diagram showing logical VMs on a mainframe
• Shows multiple VMs (VM #1, VM #2, VM #3, etc.) sharing resources on a physical mainframe

Stagnation of Virtualization (1980s-1990s)

• PC emergence reduced sharing needs due to cheap hardware
• Single-user applications dominated servers
• Server hardware was often custom-built
• Poor application isolation led to system-wide crashes

Infrastructures in the late 1990s - early 2000s

• Shows a diagram of many servers
• Each server runs a single application (e.g., File Server, Web Server, DNS Server, App Server, Domain Server)

Pre-virtualization scenario (circa 2000)

• The diagram illustrates many servers each running a single application with low resource utilization
• Only 12% to 18% of hardware is being utilized

Solution: virtualization!

• One physical server hosts multiple virtual machines (VMs)
• Each VM operates its own operating system & applications
• A hypervisor controls the physical server, and allows different VMs to safely share the same server

Post-virtualization scenario

• Significantly higher hardware utilization (55%) after implementing virtualization, compared to the earlier scenarios
• Applications still run on their own operating systems in VMs
• The hypervisor manages the VMs in a efficient way

Hypervisor scheduling

• The hypervisor manages CPU, memory, and resources among VMs.
• VMs are processed using a round-robin process.
• The hypervisor prevents virtual machines from interfering with one another

Hypervisor types

• Type-1 (bare metal) hypervisor directly communicates with the hardware. Xen, VMware ESX/ESXi, Hyper-V are examples.
• Type-2 hypervisor runs on top of a host OS and are often used for VMs on personal computers, like VMware Workstation, VirtualBox, KVM.

Why VMs instead of more processes?

• VM isolation is easier compared to multiple processes on a single OS sharing resources.
• Better security due to not sharing file system
• Isolation improves performance

Consolidation by virtualization

• Virtualization improves hardware utilization & efficiency
• Reducing energy consumption due to server consolidation

System-level virtualization

• The physical machine acts as a switch for the VMs, similar to an operating system managing processes on a CPU.
• It's a way to improve performance & efficient hardware utilization
• Guest VMs and OS are less privileged compared to the hypervisor. They have limited control over the hardware

Privilege rings (no virtualization)

• Illustrative Diagram of the different privilege levels for OS processes
• The OS kernel, device drivers, and application processes all run at different privilege levels.

De-privileging the guest OS

• The VMM manages privileged instructions
• Guest OS use the VMM to execute privileged instructions instead of directly accessing the hardware
• This ensures isolation and prevents disruption between VMs

Cloud Computing (Part II)

• The presentation was from Rodrigo Bruno
• Covered the subject of Cloud computing, part II

Recap (hypervisor!)

• Re-emphasizes Type 1 hypervisors, with examples like Xen, VMware ESX/ESXi and Hyper-V

Recap (privilege rings!)

• Explains how privileged instructions are handled in the context of virtualization
• Shows that the guest OS (users) operates at a lower privilege ring than the hypervisor's privilege ring

Recap (x86 virtualization problem!)

• Details the popf instruction as a sensitive instruction in x86 architectures that can cause problems
• Explains that the CPU will ignore popf instruction at a lower privilege ring

Recap (virtualizing virtual memory)!

• Diagram illustrating how virtual memory is handled within VMs via the hypervisor.

Brainstorming

• Focused on the topic of cloud computing and related concepts, like what cloud computing is, deployment models, service models, and Amazon Web Services

Pre-cloud enterprise infrastructures

• Describes the traditional IT approach to building computing infrastructure within a company (data centers)

Cloud Computing

• Explains the renting/subscription model of cloud computing as an alternative to building/owning a data center

Traditional IT vs. Cloud Computing Costs

• Shows a visual comparison reflecting fixed and variable costs over time, with cloud computing often having lower fixed costs but higher variable costs

Nowadays, virtually every digital-based service

• The presentation indicates that modern digital services are hosted on cloud platforms like Dropbox and Netflix. They are accessed via the internet with web interfaces or app-based interfaces.

Main Features of Cloud Computing

• Features of Cloud Computing: on-demand self-service, broad network access, shared resources pool, elastic scaling, measured services, and flexible pricing

Advantages of Cloud Computing

• Advantages of cloud computing: trade capital expenses for operational expenses, Pay-per-use, benefit from economy of scale, faster agility, and concentrate on business needs

Cloud Computing (a definition)

• Highlights the widely used NIST definition of cloud computing, emphasizing the key characteristics of cloud environments
• Defines deployment models & service models, including public/private/hybrid types & services

Deployment models

• Describes the different types of cloud deployments, such as public, private, and hybrid clouds.
• Public cloud exposes services to the general public, private cloud has internal usage only, and hybrid models use a mix of resources.

Public Cloud

• Explores the business model of public cloud providers
• Resources are shared, which leads to decreased upfront costs

Private Cloud

• Explores the business model of private cloud providers
• The cloud resources are solely available for the user or organization/company, and this often means that the user or organization is responsible for maintenance of the infrastructure

Hybrid Cloud

• Explores the business model of hybrid cloud providers
• Hybrid cloud provides a mix of public & private cloud environments based on business need or operational requirement

Service Models

• Describes the different categories of cloud services, like IaaS, PaaS, FaaS and SaaS, encompassing infrastructure as a service, platform as a service, function as a service, and software as a service, etc

CaaS (Car as a Service)

• Explores business models related to owning & leasing versus subscription based services, and relates this to cloud computing concepts to illustrate a business model example.

Cloud service models diagram

• Shows the various components involved in on-premise, IaaS, PaaS, and SaaS cloud service models
• Diagram shows relationships between elements, such as functions, applications, runtime, operating systems, networking, storage

Infrastructure as a Service (IaaS)

• Describes the infrastructure as a service (Iaas) model in the cloud.
• Diagram illustrating a high-level summary of the different layers involved in an IaaS based solution

Platfrom as a service (PaaS)

• Diagram shows the different layers of an IaaS-type solution in the cloud
• Diagram shows relationships between the different elements of the solution

Software as a Service (SaaS)

• Diagram of the services used in a SaaS-type solution in the cloud
• Diagram shows relationships between the different elements of the solution

Data Centers

• Describes various data center deployments

Inside a Data Center

• Illustrates data center architecture and physical components, including server rack configurations

Levels of Security in Data Centers

• Depicts security measures and access control layers in a data center

Sines 4.0

• Focus on energy efficiency and sustainability initiatives in the Sines region

PUE (Power Usage Effectiveness)

• Provides a quantitative measure of energy efficiency in data centers
• Explains a formula to calculate PUE, and how this can be used for efficiency comparisons

Evolution of Data Centers

• Explains how data center design has evolved over the years and different generations of designs
• Explains how energy efficiency in data centers has improved over the years

Aspects to Consider

• Outlines crucial aspects to consider concerning data centers, including 24/7 access, physical security, low electricity cost, network access, and energy and heat dissipation issues

The Importance of Electricity

• Highlights the significance of electricity costs in data centers; that the data center portion of total operational cost (10-yr TCO per rack) for a facility comprises approximately 20% of the overall cost.

Data Center Efficiency

• Diagrams show power distribution, computing loads, and overall data center energy efficiency, as measured by Useful Output / Total Power In

IT Power Load Share

• Summarizes different power components in a data center

Annual Electrical Cost of a 1MW Data Center

• Illustrates the annual electrical cost, broken down based on NCPI (Network Critical Physical Infrastructure) and IT Load Electricity consumed

Improving Data Center Efficiency

• Discusses improving efficiency through methods, such as rightsizing NCPI equipment, and using new technologies

Improvements

• Outlines different approaches to improve data center efficiency, such as using less power

Efficiency of a Typical Data Center

• Presents performance graphs of two different data centers

Energy vs. Performance

• Illustrative graph of possible tradeoffs between energy consumption in a data center and the latency/performance of the data center

Energy Carbon Intensity

• Highlights the environmental impact and offers possible ways for companies to reduce power consumption & associated carbon emissions

Cloud Computing Providers

• List of notable cloud provider companies like AWS, Microsoft Azure, Google Cloud, IBM Cloud, Alibaba Cloud, and DigitalOcean

Cloud Computing Market

• Provides market share data for various cloud providers

Information Security

• Summaries of different security methods (algorithms and models)

Security Model Based On Complete Mediation

• Diagram detailing different modules involved in security verification in a computing system

Trusted Computing Base (TCB)

• Details various guidelines for a trusted computing base (TCB) in a computer system

Real-world problem: how to exchange messages securely

• Explains the scenario for a secure information exchange process taking into account different aspects of security like authentication, integrity and confidentiality

Cryptography

• Explains cryptography as a means to securing communications over insecure networks.
• Also mentions the implementation through various algorithms for securing network communication

Cryptographic Hash

• Explains what a cryptographic hash is and what it can be used for

Cryptographic Hash (SHA1 / SHA2)

• High level details of SHA1 & SHA2 algorithms (primitives/algorithms) and their structure

Key-based Authentication (authentication + integrity) Model

• Highlights a model for securing authentication and integrity of messages exchanged (example: two parties wanting to ensure that a message originates from and has not been altered by some other party).

Confidentiality

• Details a model for encrypting messages and how this preserves confidentiality

Cryptographic Ciphers - Shared-Secret Encryption / Symmetric Encryption

• Overview of the components and structure of Symmetric encryption

Cryptographic Ciphers

• Explains the structure of Advanced Encryption Standard (AES, a symmetric cryptographic cipher), providing a high-level overview of the process, e.g., byte substitution, shift rows, mix columns.

Cryptographic Ciphers (Rivest-Shamir-Adleman)

• Provides information on the use of public key encryption for communications over insecure channels
• Illustrative diagram providing example of possible operations to encrypt & decrypt, using RSA public key encryption

Authentication + Integrity + Confidentiality

• Explains the process involved to enforce all 3 security aspects (confidentiality, integrity, and authenticity

Authentication + Integrity + Confidentiality (with hashing)

• Illustrates secure information exchange processes by including hashing to ensure integrity.

Authentication + Integrity + Confidentiality (with hashing + recv can verify)

• Illustrative diagram illustrating the concept of authentication, integrity, and confidentiality in a secure information transmission process

Fault Tolerance

• Overview of faults, errors, failures, and reliable systems and how they relate to availability & reliability

Fault Tolerance Model

• A model of classifying and handling errors within a system

Key Steps to Building Reliable Systems

• Provides crucial steps to building systems that are reliable and fault-tolerant for errors in general

Bathub Curve

• Describes the failure rate curves over time

Tolerating faults with Incremental Redundancy

• Explains how redundancy can be used to correct for the occurrence of single bit flips in network communication data (e.g., 100101 changed to 000101)

Tolerating faults with Massive Redundancy

• Explains the concept of using full replica sets of components (modules) in a system to handle module failures

N-modular redundancy

• Explains N-modular Redundancy (TMR), a method for building fault-tolerence into systems using redundancy of components / modules

Availability and Reliability

• Explanations of Availability, Downtime, Time to Failure, and the important concept of memoryless systems in the context of fault tolerance, along with mathematical definitions of the different aspects

Reliability of a TMR

• Explanations of reliability concepts and mathematical relationship to time to failures

Reliability of a TMR (long mission time)

• Illustrative graph of a reliability analysis for a TMR (Triple Modular Redundancy) system over a long period of time (e.g., 6000 hours)

Consistency

• Summarizing fault tolerance issues for a system
• Explanations of availability, fault tolerance model, incremental redundancy, and massive redundancy

State machine replication (Massive replication)

• Describes state machine replication methods to ensure consistent state among replicated systems

Active Replication

• Model for replicating states in a system

Passive replication

• Model for replicating states in a system

Consensus (Paxos)

• Explanations of Paxos protocol and how it solves consensus issues in distributed systems

Paxos

• Details the Paxos protocol used for achieving agreement within a system comprising a number of processes under possible failure modes / error conditions

CAP theorem

• Provides an overview of the CAP theorem, which highlights the trade-offs between consistency, availability, and partition tolerance in distributed systems.

Strong consistency

• Explains what strong consistency means in distributed systems in the context of how client read and write operations affect consistency.

Relaxing strong consistency - Causal Consistency

• Explains ways to enforce relaxed forms of consistency, such as Causal Consistency versus Strict Consistency

Causal Consistency

• Explanations of types of operations and how relationships between causal operations affect the results within a system

Eventual Consistency

• Explains the concept of eventually consistent systems, where updates eventually make themselves uniformly visible everywhere.