MCA Docs, MCA Docs, MCA Docs, MCA Docs, MCA Docs, MCA Docs, MCA Docs, MCA Docs, MCA Docs,, Study notes of Computer Science

Download MCA Docs, MCA Docs, MCA Docs, MCA Docs, MCA Docs, MCA Docs, MCA Docs, MCA Docs, MCA Docs, and more Study notes Computer Science in PDF only on Docsity! Cloud Cheat Sheet Term Concept Description /Analogy Illustration Real world Example Relevance to Ki Cloud Servers that are accessed over the Internet, and the software and databases that run on those servers • May be limited to a single organisation or available to multiple organisations • Users don’t have to manage physical servers themselves or run software applications on their own machines • Benefits include enhanced flexibility to cope with changing demands (scalability), potential reduced costs etc. Google cloud platform Microsoft Azure Virtual Machines A simulated, digital-only “virtual” computer that behaves as if it were a physical computer with its own hardware • VMs on the same host machine can’t interact/ access with each other • VMs make more efficient use of the hardware hosting them - by running many VMs at once, one server becomes many servers, and a data centre becomes a whole host of • data centres Windows VMs, iOS VMs, Java VMs Cloud Region A geographical area (physical location) where Cloud resources are located • Made up of at least 2-3 or more zones (depending on Cloud provider) • Each region is designed to be isolated from another to achieve the greatest fault tolerance and stability Microsoft Azure has 10 regions (e.g. Central US, East US3, West US 22) in the United States Cloud Zone One or more discrete data centres with redundant power, networking and connectivity in a Cloud region • Zones in the same region are connected through low-latency links • Spreading workload over zones better protects against issues e.g. power outages, earthquakes etc. Microsoft Azure’s Central US Region has 3 zones within Cloud Network A computer network that exists within or is part of a cloud computing infrastructure • Cloud enhances reliability by deploying redundant resources automatically when system experiences a fault • There are many ways to increase cloud reliability e.g. scaling to manage capacity surges, deploying in multiple zones etc. Google Cloud Platform network Reliability The probability of a workload performing the required functions in a period of time without failure • Involves using supervised learning techniques • Representative approaches to regression include linear regression, random forests, and neural networks Cooling failure causes outage in the Microsoft Azure data centre in the UK South region Term Concept Description /Analogy Illustration Real world Example Relevance to Ki Scalability The ability to increase or decrease IT resources as needed to meet changing demand • Used to meet the static increase in workload • Enables great flexibility that is both time- and cost-saving for businesses Using GCP, Atom bank was able to increase the number of CPUs within hours when one of its products saw a surge in demand in 2020 Vertical Scaling Moving to a larger instance or upgrading resources • Simple to do as the entire data is in a single server, so no risk of managing multiple instances simultaneously • Higher possibility of more downtime compared to horizontal scaling A website running on a 8-cpu VM that is redeployed to a 16-cpu VM Horizontal Scaling Adding more instances to a service, system or application • Cheaper than vertical scaling • Easy to do with lower chances of downtime than vertical scaling • Better resilience and fault tolerance than vertical scaling A SaaS vendor who adds instances whenever averages users per CPU (Central Processing Unit) exceeds 50 and removes instances when users per CPU falls below 40 Auto-scaling Adding instances for different purposes on demand • Performed automatically using an API (Application Programming Interface) • Enables consistent performance regardless of the current demand on resources A settlement process for a bank that horizontals scales itself depending on how many trades it’s processing Side-by-side Building scaling plans (e.g. around availability, costs) that automate how groups of different resources respond to changes in demand • Often used to temporarily create additional instances for testing or development A firm that adds development and test instances of a services as required by a project Global scaling Scaling a service to run in different geographical locations • Designed to minimise latency (i.e. speed up the delivery of content) to users in different regions through the use of a global network of proxy servers A content delivery network that delivers videos from dozens of geographically distributed data centres, so that videos are served to users from a data centre that is close to them Term Concept Description /Analogy Illustration Real world Example Relevance to Ki Shared Responsibility Model A cloud security framework that dictates the security obligations of a cloud computing provider and its users to ensure accountability • Users are responsible for their security and compliance IN the cloud • Cloud provider is responsible for the security OF the cloud • User is responsible for e.g. access management, firewall configuration, client-side data encryption • Cloud provider is responsible for e.g. protecting networking & infrastructure (regions, zones) Multi-cloud The use of multiple cloud computing and storage services in a single heterogeneous architecture • Often offers greater flexibility, reliability and more features from the “best” of each provider, compared to the use of a single provider • Also refers to the distribution of cloud assets, software, applications, etc. across several cloud-hosting environments Accenture’s IT infrastructure is 95% in the public cloud spread across Microsoft Azure, AWS and GCP Arbitrary Multi-cloud Random deployment of various workloads to different Cloud providers • Usually results from lack of Cloud strategy and poor governance • Not a planned design A firm that moves some workload into X Cloud provider then some into Y Cloud provider randomly with no plan Segmented Multi-cloud Deployment of specific types of workload to specific Cloud providers • Segregation may be determined by e.g. workload type, application type • Often results from different vendor preferences for different kind of workloads A firm deploys product A and B on X Cloud Provider, and product C on Y Cloud provider based on data classification level Choice Multi-cloud Deployment of specific workload to specific Cloud providers, with corporate governance and policy constraints to prevent random application movements and access to Cloud provider’s services • Allows users to freely use proprietary cloud services • Reduces single vendor lock-in • Bound to a single cloud providers SLA per application or service deployed A firm deploys product A on X Cloud Provider and is free to use X Cloud Provider’s services like managed databases Term Concept Description /Analogy Illustration Real world Example Relevance to Ki Parallel Multi-cloud Deployment of the same application into more than one Cloud provider • Removes the risk of a single cloud provider outage taking out service and hence guarantees a higher level of availability than what could be achieved with a single provider • Adds additional layers of complexity/abstraction and new tooling to understand, deploy and manage A firm deploys product A into both X Cloud provider and Y Cloud provider simultaneously Portable Multi-cloud Deployment of workloads anywhere and move them across different Cloud providers as you wish, anytime and anywhere • Move applications based on resource needs • Avoid vendor lock-in • Requires high levels of automation A firm deploys product A to X Cloud provider and buster to Y Cloud provider during periods of heavy utilisation