what Cloud Computing is., Study notes of Operating Systems

Download what Cloud Computing is. and more Study notes Operating Systems in PDF only on Docsity! Preliminary The following are five important characteristics that explain what Cloud Computing is: o First, you will get all your computing resources on-demand and self- service. That is, you will get Processing power, Storage and Network resources through the internet by using a simple user interface o Second, you will have the facility and infrastructure to access the above resources over the internet. o Third, these resources are provisioned from a pool through a provider, who will allocate all these resources to the customers. o Fourth, the resources are elastic, which means you can scale your resources up and down. o Fifth, you will pay for the resources that you use or reserve, which is known as the pay-for-use model. You will not be billed for the resources that you do not use. Google Cloud Platform What are the services that the Google Cloud Platform provides to its customers? Google Cloud Platform offers several services like Compute, Storage, Big Data and Machine Learning, which are used to develop applications like Web, Mobile, Analytics, or back-end solutions. Google Cloud offers reasonable billing to all the resources that it provides. The primary reasons to choose Google Cloud Platform is that it is Global, Cost-effective, Open-source friendly and Secure. Google Compute Architecture GCP offers different options for computing. Following are the options that you can choose for your requirements: RONS> You can work in a serverless environment You can use a Managed application platform You can use Container Technology You can build your own cloud-based infrastructure for more control and flexibility GCP Region, Zones and Pricing GCP is organized into Regions and Zones. For example, if you want to launch a Virtual Machine (VM) in GCP using Compute Engine (briefly described in later topics), it will run in a zone that you specify. A collection of zones is known as Region. All the zones within a region will have faster connectivity among them. You can also distribute your resources across different zones in a region. As of now, GCP provides 20 Regions and 61 Zones to its customers. Customers duplicate their application across multiple regions to prevent application loss. Google provides Per Second Billing for a VM. It also provides billing for various services like Kubernetes Engine, which is Container as a Service (CaaS), Cloud DataProc, which is a Big Data system (Hadoop as a Service), and App Engine, which is Platform as a Service (PaaS). Security The following are the solutions that Google Cloud provides to its customers to keep their data safe: Google will design its own Server Boards and Network Equipment. Cryptographic signatures are used by Google server machines to check that they are booting the correct software. Google will design and build its own data centers by introducing multiple layers of physical security protections. Google Infrastructure provides the cryptographic privacy and integrity for data-on-the-network. Google also has Encryption support in hard drives and SSDs. Google services that are made available through the internet should register with an infrastructure service called Google Frontend, whose job is to check for incoming network connections for correct certificates. Google also supports multi-layer, multi-tier denials of service protections. Phishing attacks against Google employees are guarded through U2F Compatability security keys. GCP Resource Hierarchy The gathering of Resources will make up the GCP Services. Resources can be Compute Engine Virtual Machine, Cloud Pub/Sub, Cloud Storage Bucket, App Engine and so on. These resources are organized into Projects. Projects can be combined into Folders, and the folders can contain Sub Folders. All the folders and projects are organized under Organization Node. Policies are defined in Projects, Folders and Organization nodes. Policies are inherited downwards of Resource Hierarchy. GCP Resource Hierarchy In the above image, you have a VPC network with a subnet, which consists of two Compute Engine VMs. Even though both the VMs are in different zones, they are considered as neighbors as they are in the same subnet. More About Google Compute Engine The basic functionality of the Compute Engine is to create and run the Virtual Machines on the Google Infrastructure. You can create your own Virtual Machine in GCP by using GCP Console or gcloud command-line tool. Your VM is capable of running images of either Linux or Windows provided by Google or you can create your own Custom Images or you can import images from your Physical Servers. You can choose how much Memory and how many Virtual CPUs your VM should have. Regarding Processing Power, if you have workloads like Machine Learning and Data Processing, then you can consider GPUs for your VM. More About Google Compute Engine Persistent Storage, that is made available to your VM is of two types, namely, Standard and SSD. You can also add Local SSD for high performance. The data storage in Local SSD is not safe because when the VM terminates, data will get erased. Therefore, it is better to consider Persistent Disks for the permanent storage of your data. The default disk is Standard Persistent Disk If you want your VM to spin up with configurations like Installing software packages on first boot, you can do it with Start-up scripts. There are also other fields like Metadata, which gives unique information about the VM. You can take snapshots of your disk for the backups. Suppose if you want to do a batch job like analyzing large data sets, then consider options like Preemptible VM, which saves a lot of money. The maximum number of virtual CPUs in a VM is 96, and the maximum memory size is 624 Gigabytes, which will be useful for heavy workloads like In-memory databases and CPU intensive analytics. GCP also supports a feature called AutoScaling, which allows you to add or remove VMs. You can also balance the incoming traffic among different VMs through Load Balancing. Google Cloud Storage and Bigtable Introduction Developers develop many applications that require the storage of large amounts of data. Data can be in many forms like Media, Confidential data from devices, Customer account balances, and so on. Earlier, we read that data can be stored in Persistent disks. Also, GCP provides other storage options for Structured or Unstructured data, Transactional, and Relational data. More About Cloud Storage Google Cloud Storage is an Object Storage. Google Cloud Storage will store the data in objects, which will be in the form of an arbitrary bunch of bytes, that can be addressed using a unique key. These unique keys are in the form of URL, which makes it easier to interact with Web Technologies. Cloud storage is comprised of Buckets, which is used to hold the Storage Objects. Storage Objects are immutable, that is, for every new change, a new version will be created. Object Versioning is another important concept in cloud storage. If you turn ON this feature, it will store every version of the object. Otherwise, the newer version will override the old one. When you create a bucket, you will provide a Unique name to it, specify a Geographic location to store the data and also choose a Storage Class by default. You can control the access to the bucket using Cloud IAM and Access Control List (ACL). About Cloud Bigtable The following are a few reasons to select Cloud Bigtable as your storage option: Scalability: With the help of Bigtable, you can increase your machine count, which does not require any downtime. Also, it handles all Administration tasks like Upgrades and Restarts. The data present in the Cloud Bigtable is encrypted. You can use IAM roles to specify who can access the data. From an application API perspective, data is written to and from Bigtable through data service layers like Managed VMs, HBase Rest Server, or Java Server using HBase Client. Cloud Bigtable will serve the data to Applications, Dashboards and Data services. Data can be read from and written to Bigtable through batch processes like Hadoop MapReduce, DataFlow or Spark.