Google Cloud Associate Engineer

Static IP Addresses

Are assigned for extended periods of time.

Ephemeral IP Addresses

Are attached to VMs and released when the VM is stopped.

Persistent Disk SSDs

Are often used for low-latency applications where persistent disk performance is important.. SSDs cost more than HDDs.

Multiregion Storage

Provides for storing replicas of objects in multiple Google Cloud regions, which is important for high availability, durability, and low latency. Allows for faster access to data when users or applications are distributed across regions.

Cloud Storage

Cloud Storage is GCP's object storage system. Objects can be any type of file or binary large object. Objects are organized into buckets. It is important to remember that buckets share a global namespace, so each bucket name must be globally unique. Cloud Storage is not a file system, it is a service that receives, stores, and retrieves files or objects from a distributed storage system. Cloud Storage is not part of a VM in the way an attached persistent disk is. Cloud Storage is accessible from a VM or any other network device with appropriate privileges. Each stored object in Cloud Storage is uniquely addressable by a URL GCP users and other can be granted permission to read and write objects to a bucket. Cloud Storage is useful for storing objects that are treated as single units of data. For example, an image file is a good candidate for object storage. If you write or retrieve an object all at once and you need to store it independently of servers that may or may not be running at any time, then Cloud Storage is a good option. There are different classes of Cloud Storage: Regional, Multiregional, Nearline, and Coldline Both regional and multiregional storage are used for frequently used data.

Use Cases for Compute Engine Virtual Machines

Compute Engine is a good option for when you need maximum control over VM instances. With Compute Engine you can do the following: - Choose the specific image to run on the instance - Install software packages or custom libraries - Have fine-grained control over which users have permissions on the instance - Have control over SSL certificates and firewall rules for the instance - Provides the least amount of management relative to other computing services in GCP.

Constraints of Resources

Constraints are restrictions on services. GCP has list constraints and Boolean constraints. List constraints are lists of values that are allowed or disallowed for a resource. For example: -Allow a specific set of values -Deny a specific set of values -Deny a value and all its child values -Allow all allowed values -Deny all values Boolean constraints evaluate true or false statements and determine whether the constraint is applied or not. For example, if you want to deny access to serial ports on VMs, you can set constraints/compute.disableSerialPortAccess to True.

Cloud Load Balancing

Google provides global load balancing to distribute workloads across your cloud infrastructure. Using a single multicast IP address, Cloud Load Balancing can distribute the workload within and across regions, adapt to failed or degraded servers, and autoscale your compute resources to accommodate changes in workload. Cloud Load Balancing also supports internal load balancing, so no IP addresses need to be exposed to the Internet to get the advantages of load balancing. Cloud Load Balancing can load-balance HTTP, HTTPS, TCP/SSL, and UDP traffic.

Cloud Spanner

Google's globally distributed relational database that combines the key benefits or relational databases, such as strong consistency and transactions, with the ability to scale horizontally like a NoSQL database. Spanner is a highly available database with a 99.999 SLA making it a good option for enterprise applications that demand scalable, highly available relational database services. Cloud Spanner also has enterprise-grade security with encryption at rest and encryption in transit, along with identity-based access controls. Cloud Spanner supports ANSI 2011 standard SQL.

Folder

Organizations contain folders and folders contain other folders or other projects. A single folder may contain both folders and projects.

Virtual Private Cloud (VPC)

Your internal GCP network. Here you specify IP addresses for your VMs and services and define firewall rules to control access to to subnetworks and VMs in your VPC. A VPC can span the globe without relying on the public internet. VPCs can have subnets in any GCP regions world-wide and subnets can span the zones that make up a region. You can have resources in different zones on the same subnet.

Billing Budgets and Alerts

The GCP billing service includes an option for defining a budget and setting billing alerts. Budgets are associated with billing accounts and not projects. One or more projects can be linked to a billing account, so budgets and alerts should be based on what you expect to spend on all projects. With a budget you can set three alert percentages or more. When a percentage of a budget is spend billing administrators and billing account users will be notified by email.

Latency

The amount of time it takes to retrieve data.

Principle of Least Privilege

The practice of assigning permissions that are needed and no more is known as the principle of least privilege and it is one of the fundamental best practices in information security.

Organization

The root of the resource hierarchy and typically corresponds to a company or organization. G-suite domains and Cloud Identity accounts map to GCP organizations. If your company does not use G-suite, you can use Cloud Identity, Google's Identity as a Service (IDaaS) offering. When an Organization is created, all users in that organization are granted Project Creator and Billing Account Creator roles.

Organization Administrator

Users who are responsible for: -Defining the structure of the resource hierarchy -Defining identity access management policies over the resource hierarchy. -Delegating other management roles to other users.

Compute Engine Security Admin

Users with this role can create, modify, and delete SSL certificates and firewall rules.

Compute Engine Network Admin

Users with this role can create, modify, and delete most networking resources, and have read-only access to firewall rules and SSL certifications. This role does not give the user permission to create or alter instances.

Compute Engine Viewer

Users with this role can get and list Compute Engine resources but cannot read data from those resources.

Compute Engine Admin

Users with this role have full control over Compute Engine instances.

Factors to consider when running VMs in Zones & Regions

- Cost, which can vary by region - Data locality regulations, such as keeping data about EU citizens in the EU. - High availability, if you are running multiple instances, you may want them in different zones and possibly different regions. - Latency, keeping instances and data geographically close to users can help reduce latency. - Need for specific hardware platforms, which can vary by region.

Limitations of Preemptible Virtual Machines

- May terminate at any time, if they terminate within 10 minutes of starting, you will not be charged for that time. - Will be terminated within 24 hours. - May not always be available, availability may vary across zones and regions. - Cannot migrate to a regular VM - Cannot be set to automatically restart - Are not covered by any service level agreement (SLA)

Identities

Are abstractions about users of services, such as human user. After an identity is authenticated by logging in the authenticated user can access resources and perform operations based on the privileges granted to that identity.

Cloud Firestore

A GCP-managed NoSQL database service designed as a backend for highly scalable web and mobile applications. One advantage of Cloud Firestore is that it is designed for storing, synchronizing, and querying data across distributed applications, like mobile apps. Apps can be automatically updated in close to real time when data is changed on the backend. Cloud Firestore supports transactions and provides multiregional replication.

Coldline Storage

A low-cost archival storage designed for high durability and infrequent access. Suitable for data that is accessed less than once per year and has at least a 90-day minimum storage. There are costs associated for retrieving data stored in Coldline Storage.

Cloud Datastore

A NoSQL document database. This database uses the concept of a document, or collection of key-value pairs, as a basic building block. Documents allow for flexible schemas. For example, a document about a book may have key-value pairs listing author, title, and date of publication. Cloud Datastore is accessed via a REST API that can be used from applications running in Compute Engine, Kubernetes Engine, or App Engine. This databases will scale automatically based on load. It will also shard, or partition data as needed to maintain performance. Datastore is a managed service, that takes care of replication, backups, and other database administration tasks. Cloud Datastore is well suited for applications that demand high scalability and structured data and do not always need strong consistency when reading data. Datastore is used for nonanalytic, non relational storage needs. Product catalogs, user profiles, and user navigation history are examples of the kinds of applications that use Cloud Datastore.

Virtual Machines

A basic unit of computing resources. GCP offers preconfigured VMs with varying numbers of vCPUs and amounts of memory. You can create a custom configuration if the preconfigured offerings don't meet your needs. You can create multiple VMs running on different OS and applications. VMs are abstractions of physical servers, they are essentially programs that emulate physical servers and provide CPU, memory, storage, and other services.

Cloud SDK

A command-line interface for managing GCP resources, including VMs, disk storage, networking firewalls, and virtually any other resource you might deploy in GCP. Cloud SDK has client libraries for Java, Python, Node.js, Ruby, GO, .NET, and PHP. The Cloud SDK is available as a Docker Image which is a really easy and clean way to work with it.

Infrastructure as a Service (IaaS)

A computing service where customers can create and manage VMs themselves. This model gives the cloud user the greatest control of all the computing services. Users can choose the OS to run, which packages to install, and when to backup and perform other maintenance operations. Compute Engine is GCP's IaaS product.

Cloud DNS

A domain name service provided in GCP. Cloud DNS is aa high availability, low-latency, service for mapping from domain names, such as example.com to IP addresses. Designed to automatically scale so customers can have thousands and millions of addresses without concern for scaling the underlying infrastructure. Also provides for private zones that allow you to create custom names for your VMs if you need those.

Region

A geographical location, such as asia-east1, europe-west2, and us-east4. The zones within a region are linked by low-latency, high-bandwidth network connections.

Nearline Storage

A good option for when data needs to be kept for extended periods of time but is rarely accessed. It costs less than regional or multiregional storage and is optimized for infrequent access. There are costs with retrieving data stored in Nearline Storage. Nearline storage is designed for use cases in which you expect to access files less than once per month.

Internal IP Addresses

Are accessible to only services in your internal GCP network.

Apigee API Platform

A management service for GCP customers providing API access to their applications. Allows developers to deploy, monitor, and secure their APIs. It also generates API proxies baed on the Open API specification. It is difficult to predict load on an API, and sometimes spikes can occur. For those times, the Apigee API platform provides routing and rate-limiting based on policies customers can define. APIs can be authenticated using either OAuth 2.0 or SAML. Data is encrypted both in transit and at rest in the Apigee API platform.

Cloud Armor

A network security service that gives you the ability to allow or restrict access based on IP address, predefined rules to counter cross-site scripting attacks, ability to counter SQL injection attacks, ability to define rules at both level 3 (network) and level 7 (application). It also allows and restricts access based on the geolocation of incoming traffic.

Cloud Functions

A platform for running code in response to an event such as uploading a file to Cloud Storage or adding a message to the message queue. Cloud Functions work well when you need to respond to an event by running a short process coded in a function or by calling a longer-running application that might be running on a VM, managed cluster, or App Engine. This computing service is not designed to execute long-running code. Cloud Functions will automatically scale as load increases. Cloud Functions is often used to call other services, such as a third-party API or other GCP services, like natural language translation.

Roles

A role is a collection of permissions. Roles are granted to users by binding a user to a role. Permissions cannot be assigned to users, they can be assigned only to roles that have permissions. There are three types of roles in GCP: - Primitive Roles - Predefined Roles - Custom Roles

Compute Engine

A service that allows users to create VMs, attached persistent storage to those VMs, and make use of other GCP services like Cloud Storage. There are discounts applied when a VM is run for more than 25% of the month.

Stackdriver

A service that collects metrics, logs, and event data from applications and infrastructure and integrates the data so DevOps engineers can monitor, assess, and diagnose operational problems.

Cloud Interconnect

A set of GCP services for connecting your existing networks to the Google network. Cloud Interconnect offers two types of connections: Interconnect and Peering

Object Storage

A system that manages the use of storage in terms of objects or blobs. Usually these objects are files that are grouped into buckets. Each object is individually addressable by a URL. Object Storage is not limited by the size of disks or SSDs attached to a server Objects can be uploaded with out concern for the amount of space available on a disk. Multiple copies of objects are stored to improve availability and durability. In some cases copies of objects may be stored in different regions to ensure availability even if a region becomes inaccessible. Object Storage is serverless, there is no need to attach VMs and attach storage to them. GCP's Object Storage is called Cloud Storage and is accessible from servers running on GCP and other devices with internet access. Access controls can be applied at the object level, this allows users of Cloud Storage to control which users can access and update objects. It takes longer to. retrieve data from object storage than it does from block storage.

Zone

A zone is a data center that may be compromised of one or more closely coupled data centers. Zones are located within regions.

Custom Roles

Allow cloud administrators to create and administer their own roles. Custom roles can only be used at the project or organization levels and not the folder level. Custom roles are assembled using permissions defined in IAM. While you can use most permissions in a custom role, some are not available in custom roles.

Serverless Computing

Allows developers and application administrators to run their code in a computing environment that does not require setting up VMs or Kubernetes Clusters. Serverless Computing options in GCP include App Engine and Cloud Functions.

Cloud CDN

Allows users anywhere to request content from systems distributed in various regions. CDNs enable low-latency response to these requests by caching content on a set of endpoints across the globe. CDNs are especially important for sites with large amounts of static content and a global audience. News sites, for example, could use the Cloud CDN service to ensure fast response to requests from any point in the world.

Peering

Allows you to share data and network access between an on-premise data center and your VPC. There are several types of peering available with your VPC in GCP.

Cloud Memorystore

An in-memory cache service. A managed Redis service for caching frequently used data in memory. Cloud Memorystore allows users to specify the size of a cache while leaving administration tasks to Google. GCP ensures high availability, patching, and automatic failover.

App Engine Structure

App Engine applications consist of services. Services provide a specific function, like computing sales tax in a retail web application or updating inventory as products are sold on a site. Services have versions and this allows multiple versions to run at one time. Each version of a service runs on an instance that is managed by App Engine. The number of instances used to provide an application depends on your configuration for the application and the current load on the application. Autoscaling is possible with dynamic instances. Resident instances on App Engine run continually and can be added/removed manually. GCP allows users to setup daily spending limits as well as create budgets and set alarms for costs.

Persistent Disk HDDs

Applications that require large amounts of persistent disk storage but can tolerate longer read and write times can use HDDs to meet their storage requirements.

Caches

Caches are in-memory data stores that maintain fast access to data. The latency of im-memory data stores is designed to be submillisecond. Caches are quite helpful when you need to keep read latency to a minimum in your application. Memory in a cache is more expensive than SSD or HDD storage. Caches are volatile, you lose the data stored in the cache when power is lost or the OS is rebooted. A cache should never be used as the only data source for storing data, some form of persistent storage should be used to maintain a data store that always has the latest and most accurate version of the data. Caches can get out of sync with the system of truth, this happens when the system of truth is updated but the new data is not written to the cache.

Autoscalers

Can add/remove VMs from a cluster based on the workload, this is called autoscaling. This helps control costs by not running more VMs than needed and also ensures that sufficient computing capacity is available when workloads increase.

Cloud Storage: Lifecycle Management

Can automatically manage objects based on policies you define. For example, you could define a policy that moves all objects more than 60 days old in a bucket to Nearline Storage or delete any object in a Coldline Storage bucket that is older than 5 years. Lifecycle management policies are applied to buckets and affect all objects in the bucket. You can delete an object or change its storage class. Both unversioned and versioned objects can be deleted. If the live version of a file is deleted, then instead of actually deleting it, the object is archived. If an archived version of an object is deleted, the objects is permanently deleted. Multiregional and regional storage objects can be changed to nearline or coldline. Nearline can be changed only to coldline.

Firewall Rules

Can be configured to to limit inbound and outbound traffic to the IP address of the application server or load balancer in front of the application cluster.

External IP Addresses

Can be either static or ephemeral. Static addresses are assigned to a device for extended periods of time. Ephemeral external IP addresses are attached to VMs and released when the VM is stopped.

Specialized Services

Can be used as building blocks of applications or as part of a workflow for processing data. Specialized services commonly are serverless, provide a specific function such as translating text or analyzing images, and provide an API to access the functionality of the service. Some of the specialized services in GCP are: AutoML - a machine learning service Cloud Natural Language - a service for analyzing text Cloud Vision - a service for analyzing images Specialized services encapsulate advanced computing capabilities and make them accessible to developers who are not experts in the domains provided.

Billing Account Creator

Can create new self-service billing accounts.

Containers

Containers are like lightweight VMs that isolate processes running in one container from processes running in another container on the same server. Containers are good options when you need to run applications that depend on multiple micro services running in your environment. The services are deployed through containers and GCP takes care of monitoring, networking, and some security managements tasks. Containers can start and stop in seconds and use fewer resources in comparison to VMs.

Container Manager

Coordinates containers running on the same server within Kubernetes Clusters. Ensures isolation between running containers.

Custom Images

Custom Images are especially useful if you have to configure an operating system and install additional software on each instance of a VM that you run. Instead of repeatedly configuring and installing software for each instance, you could configure and install once and then create a custom image from the boot disk of the instance.

Custom Machine Types

Custom machine types can have between 1 and 96 vCPU and up to 6.5 GB of memory per vCPU. The price of a custom machine type is based on the number of vCPUs and the memory allocated.

Cloud Bigtable

Designed for petabyte-scale applications that can manage up to billions of rows and thousands of columns. It is based on a NoSQL model known as wide-column data model. Wide-column databases, as the name implies, store tables that can have a large number of columns. Not all rows need to use all columns, so in that way iit is like Datastore -- neither require a fixed schema to structure the data. Bigtable is suited for applications that require low-latency write and read operations. It is designed to support millions of operations per second. Bigtable integrates with other Google Cloud services, such as Cloud Storage, Cloud Pub/Sub, Cloud Dataflow, and Cloud Dataproc. It also supports the Hbase API, which is an API for data access in the Hadoop big data ecosystem. Bigtable also integrates with open source tools for data processing, graph analysis, and time-series analysis. Bigtable runs in clusters and scales horizontally. Bigtable is designed for applications with high data volumes and a high-velocity ingest of data. Time series, IoT, and financial applications all fall into this category.

Networking

Each network-accessible device or service in your device will need an IP address. Devices within GCP can have both internal and external addresses.

Billing Account User

Enables a user to link projects to billing accounts.

Billing Account Viewer

Enables a user to view billing account costs and transactions.

Ephemeral Disk

Exist and store data only as long as the VM is running. Store OS files and other files and data that are deleted when the VM is shut down.

File Storage

File Storage services provide a hierarchical storage system for files. GCP has a file storage service called Cloud Filestore. File Storage is suitable for applications that require operating system-like file access to files. The file storage system decouples the file system from specific VMs. The file system, its directories, and its files exist independent of VMs or applications.

Data Analytics

GCP has a number of services designed for analyzing big data in batch and streaming modes: BigQuery - A petabyte-scale analytics database service for data warehousing. Cloud Dataflow - A framework for defining batch and stream processing pipelines. Cloud Dataproc - A managed Hadoop and Spark service. Cloud Dataprep - A service that allows analysts to explore and prepare data for analysis.

Organization Policies

GCP provides an Organization Policy Service. This service controls access to an organization's resources. The Organization Policy Service lets you specify limits on the ways resources can be used. Organization policies are defined in terms of constraints on a resource.

Databases

GCP provides several database options, some are relational databases and some are NoSQL databases. Some are serverless and others require users to manage clusters of servers. GCP users must understand their application requirements before choosing a service, and this is especially important when choosing a database.

Identity Management

GCP's IAM service enables customers to define fine-grained access controls on resources in the cloud. IAM uses the concepts of users, roles, and privileges.

Cloud SQL

GCP's managed relational database service that allows users to setup MySQL or PostgreSQL databases on VMs without having to attend to database administration tasks, such as backing up a database or patching database software. Includes management of replication and allows for automatic failover, providing for highly available databases. Relational databases are well suited to applications with relatively consistent data structure requirements . For example, a banking database that may track account numbers, customer names, addresses, and so on..

Partner Interconnect

If an organization cannot achieve a direct interconnect with a Google facility, it could use Partner Interconnect. This service depends on a third-party network provider to provide connectivity between the company's data center and a Google facility.

App Engine: Flexible Environment

In the flexible environment, you run Docker containers in the App Engine environment. With Docker files you can specify a base OS image, additional libraries and tools, and custom tools. The flexible environment works well in cases where you have application code but also need libraries or other third-party software installed. It is a good option when you can package your application and services into a small set of containers. These containers can be autoscaled according to load. As the name implies, the flexible environment gives you more options, including the ability to work with background processes and write to local disk. There will always be at least one container running with your service, and you will be charged for that time even if there is no load on the system.

App Engine: Standard Environment

In the standard environment, you run applications in a language-specific sandbox, so your application is isolated from the underlying servers OS as well as other applications running on that server. The standard environment is well suited to applications that are written in one of the supported languages and do not need OS packages or other compiled software that would have to be installed along with the application code. Supported Languages are Java, Python, PHP, Node.js, and Go. There are no running instances when there is no load.

Primitive Roles

Include Owner, Editor, and Viewer. These are the basic privileges that can be applied to most resources. It is a best practice to use predefined roles instead of primitive roles when possible. Primitive roles grant a wide range of permissions that may not always be needed by a user.

Virtual Machine Images

Instances run on images, which contain operating systems, libraries, and other code. You may choose to run a public image provided by Google, both Windows and Linux images are available. You can also run public images provided by open-source projects or third-party vendors. You can also create a custom image from a boot disk or by starting with another image in the even that a public image does not meet your needs.

Kubernetes Engine

Is designed to allow users to easily run containerized applications on a cluster of servers (VMs). With containers, processes and resources are isolated using features of the host OS. With this approach, there is no need for a hypervisor as the host OS maintains isolation. A Container Manager is used and it coordinates containers running on the server. No additional, or guest OSs run on top of the container manager. Containers make use of the host OS functionality, while the OS and Container Manager ensure isolation between running containers. Kubernetes Engine allows users to describe the compute, storage, and memory resources they'd like to run their services. Kubernetes Engine then provisions the underlying resources. It's easy to add/remove resources from a Kubernetes Cluster using a command-line interface or graphical user interface. With Kubernetes, you can administer the cluster, specify policies such as autoscaling, and monitor cluster health.

Regional Storage

Keeps copies of objects in a single cloud region. Regional storage is well suited for applications that run in the same region and need low latency access to objects in Cloud Storage.

Managed Kubernetes Clusters

Managed Clusters make use of containers. In a managed cluster you can specify the number of servers you'd like to run and the containers that should run on them. You can also specify autoscaling parameters to optimize the number of containers running. In a managed cluster, the health of containers is monitored for you. If a cluster fails GCP will detect it and start another cluster for you.

Billing Account Administrator

Manages billing accounts but cannot create them.

Cloud Storage for Firebase

Mobile app developers may find Cloud Storage for Firebase to be the best combination of cloud object storage and the ability to support uploads and downloads from mobile devices with sometimes unreliable network connections. The Cloud Storage for Firebase API is designed to provide secure transmission as well as robust recovery mechanisms to handle potentially problematic network quality. Once files like photos or music recordings, are uploaded into Cloud Storage, you can access those files through the Cloud Storage command-line interface and software development kits (SDKs).

Persistent Disk

Persistent Disks are a storage service that are attached to VMs in Compute Engine or Kubernetes Engine. Provide Block Storage on SSDs and HDDs. An advantage of persistent disks on GCP is that these disks support multiple readers without a degradation in performance. This allows for multiple instances to read a single copy of data. Disks can also be resized as needed while in use without the need to restart your VMs. Continues to exist and store data even if it's detached from a virtual server or the server to which it is attached shuts down. Are used when you want data to exist on a block storage device independent of the VM. These disks are good options when you have data that you want available independent of the lifecycle of the VM, and support for fast OS system and file system access.

Policy Evaluation

Policies are inherited and cannot be disabled or overriden by objects lower in the hierarchy. Multiple policies can be in effect for a folder or project.

Project

Projects are the basis for enabling and using GCP services like managing APIs, enabling billing and adding and removing collaborators and enabling other Google services. Each project is a separate compartment and each resource belongs to exactly one. Anyone with the resourcemanager.projects.create IAM permission can create a project. Your organization will have a quota of projects it can create. The quota varies based on typical use, the customer's usage history, and other factors decided by Google.

Load Balancers

Provide a single access point to a distributed backend. This is useful when you need to have high availability for your application. If one VM in your cluster fails, the workload can be directed to another VM in the cluster.

Predefined Roles

Provide granular access to resources in GCP and they are specific in GCP products. By using predefined roles you can grant only the permissions a user needs to perform their function. Predefined roles are grouped by service.

Platform as a Service (PaaS)

Provides a runtime environment to execute applications without the need to manage underlying servers, networks, and storage systems. App Engine and Cloud Functions are GCP's PaaS offerings.

Cloud Filestore

Provides a shared file system for use with Compute Engine and Kubernetes Engine. Filestore can provide high numbers of input-output operations per second (IOPS) as well as variable storage capacity. File system administrators can configure Cloud Filestore to meet their specific IOPS and capacity requirements. Filestore implements the Network File System (NFS) protocol so system administrators can easily mount shared file systems on virtual servers.

Service Account

Service Accounts are given to applications or VMs so that they can act on behalf of a user or perform operations that the user does not have permissions to perform. Services accounts are sometimes treated as resources and also identities. There are two types of service accounts, user-managed service accounts and Google-managed service accounts. User can create up to 100 service accounts per project. When you create a project that has the Compute Engine API enabled, a Compute Engine service account is created automatically. The Compute Engine service account will be granted editor role in the project where it is created. Service accounts can be managed as a group of accounts at the project level or at the individual service account level. Service accounts use an email and cryptographic keys. You can change the permissions of Service Accounts without having to recreate a VM.

Preemptible Virtual Machines

Short-lived compute instances suitable for running workloads for applications that perform financial modeling, rendering, big data, continuous integration, and web crawling options. These VMs can persist for up to 24 hours. If an application is fault-tolerant and can withstand possible instance interruptions then using preemptible VMs can help reduce Google Compute Engine costs significantly.

Billing Accounts

Store information about how to pay charges for resources used. A billing account is associated with one or more projects. All projects must have billing accounts unless they only use free services. You may have one or multiple billing accounts. There are two types of billing accounts: - Self-serve: Are paid by credit card or direct debit from a bank account, the costs are charged automatically. - Invoice: Bills or invoices are sent to customers. This type of account is commonly used by enterprises or large customers.

Block Storage

Uses a fixed-size data structure called a block to organize data. Block storage is commonly used in ephemeral and persistent disks attached to VMs. With a block storage system, you can install file systems on top of the block storage, or you can run applications that access blocks directly. Some relational databases(RDB) can be designed to access blocks directly rather than working through file systems. RDB's often write directly to blocks. Block storage is available on disks that are attached to VMs in GCP. Block storage can either be persistent or ephemeral. A persistent disk It takes longer to. retrieve data from object storage than it does from block storage.

Billing

Using resources such as VM's, object storage, and specialized services usually incurs charges. The GCP Billing API provides a way for you to manage how you pay for resources used.

Interconnect

Utilizes direct access to networks using Address Allocation for Private Internets standard to connect to devices in your VPC. A direct connection is maintained between an on-premise or hosted data center and one of Google's colocation facilities.

Creating a VM

VMs come in a range of predefined sizes, but you can also create a customized configuration. When you create a VM instance you can specify the number of parameters including the following: -The OS -Size of persistent storage -Adding graphical processing units (GPUs) for compute-intensive operations like machine learning (ML) -Making the VM preemptible

Hypervisor

VMs run within a low-level service called hypervisor. GCP uses a security hardened version of the KVM hypervisor. KVM stands for Kernel Virtual Machine and provides virtualization on Linux systems running on x86 hardware. Hypervisors run on an OS like Linux or Windows Server. Hypervisors can run multiple OS while keeping the activities of each isolated from other guest OSs. Each instance of an executing guest OS is a VM instance.

Peered Interconnect

VPN services that enable traffic to transmit between data centers and Google facilities using the public internet.

App Engine

With App Engine, developers and application administrators don't need to concern themselves with configuring VMs or specifying Kubernetes clusters. Developers create applications in a popular programming language and deploy that code to a serverless applications. App Engine manages the underlying computing and network infrastructure, there is no need to configure VMs or harden networks to protect your application. Is used for applications and containers that run for extended periods of time, such as a website backend, point-of-sale system, or custom business applications. App Engine is available in two types: standard and flexible.

Exporting Billing Data

You can export billing data for later analysis or for compliance reasons. Billing data can be exported to either a BigQuery database or a Cloud Storage file. When exporting to a file, you will need to specify a bucket name and report prefix. You have the option of choosing either the CSV or JSON file format.

Compute Engine Snapshots

You can take a snapshot of a Compute Engine persistent disk to quickly back up the disk so you can recover lost data, transfer contents to a new disk, or make static data available to multiple nodes. Snapshots can be used as an image for other VMs.