When discussing the components of a cloud-based service, we talk about “cloud infrastructure.” It has a different system architecture and method of providing services compared to on-premise data centers. Most cloud computing infrastructures are hosted by a third party and made available via the internet. Virtualizing and abstracting hardware resources enables resource scaling, sharing, and provisioning for users situated in different physical locations.
Cloud suppliers can consequently market computational functionality as a service to end-users that do not own, administer, and run the IT infrastructure in the same way as their on-site data centers.
Components, characteristics, service, and deployment models for cloud infrastructure will be discussed.
Components of Cloud Infrastructure:
A cloud service comprises client-side systems such as PCs, tablets, and other devices connected to the network’s backend data center components. The components that compose a cloud infrastructure include:
Network:
The network acts as a conduit for data transmission between remote servers in the cloud and local clients. A cloud data center is an off-site facility where computer tasks are executed. Users access and interact with these components across private or public networks that transmit data between the two ends of a cloud service. The data is the information transmitted via the network, which might be anything from images to logs to commands.
According to the Open Systems Interconnection (OSI) data communications model, the network comprises the physical and electrical components like routers, cables, and switches, as well as the software applications and hardware firmware that enable data transfer.
Hardware:
A group of virtual hosts, standing in for a predetermined group of physical hardware parts, is used to gain access to cloud computing. A variety of hardware assets common to any data center, whether in the cloud or on-premise, lie beneath the layers of abstraction and software-defined architecture, over which end users have no control, management, or operation. Servers, CPUs, GPUs, power supplies, RAM, and other hardware components are included.
Depending on the cloud service model, virtualization and abstraction layers can be used to scale the allocation of these physical resources between users and IT workloads. The hardware solutions are designed with redundancy and flexibility to prevent performance, security, and availability issues from affecting end-users in the cloud.
Storage:
The cloud infrastructure stack would be incomplete without the platform and storage system. Cloud data centers use a wide range of storage media to protect data, retain copies, and evenly distribute storage space among users. Virtualization or a software-defined architecture abstracts the storage infrastructure’s underlying hardware stack. Users can take advantage of storage as a cloud service, adding or removing it as needed without having to deploy hardware on each server.
Formats commonly used for cloud storage include:
Block Storage:
With this method, data is broken up into smaller chunks that may be distributed across numerous server farms’ worth of storage devices. The information is isolated from the physical infrastructure it resides on. Separate blocks can be created from a single storage volume. Data assets that don’t change often are best suited for block storage.
Object Storage:
Uncompressed and unencrypted data items are broken off from larger files, assigned their own metadata identifiers, and then saved. Metadata can be altered to suit individual needs (Whereas with block storage, only a small subset of metadata features can be used as unique identifiers, this is not the case with key-value stores). Object storage is the best option when dealing with data assets in a constant state of flux.
File Storage:
This is related to NAS, or network access storage, and functions like the hard drive or another storage medium on your computer. The configuration is simple, and it only uses one data path.
Virtualization:
The cloud service is separated from its hardware resources, such as processing power and storage, utilizing virtualization or other software-defined computing architecture. The platform, processing, storage, and networking capabilities of physical hardware are simulated in software so that users can act as though they had access to the corresponding virtual resources.
Cloud suppliers operate and control the hardware resources that support a cloud service. In this model, consumers only pay for the services they really use, which means that the flaws in hardware supporting a cloud service must not affect the Service Level Agreement (SLA) (SLA). IT workloads can be dynamically relocated and allocated among a pool of hardware resources in virtualized and reconfigurable IT environments, hiding these constraints from cloud service consumers.
Cloud Infrastructure:
Due in huge part to the nature of cloud service operations and the specific architecture required for cloud computing, cloud infrastructure has distinct features that set it apart from traditional on-premises data centers. Some of the features of clouds are as follows:
- Extensible to a large size
- Sharing of resources in a flexible fashion
- Autonomous service provisioning on demand
- Safe against hacking thanks to several safeguards working in tandem
- Delivery model based on the usage of the service, or “as a service.”
- An abundance of readily available information technology resources and services
- Provider-managed in the cloud
Cloud Infrastructure Deployment Models:
A single user can use infrastructure in the cloud with restricted access or share it among several users. No matter the deployment model, the underlying infrastructure resources are the same but are distributed differently.
The three primary types of cloud deployment methodologies are as follows:
Public Cloud:
A collection of virtualized resources is shared by many users outside the vendor’s firewall. The service is provided on an as-needed basis and is billed using a pay-per-use billing mechanism. The vendor is accountable for managing and operating the public cloud.
Private Cloud:
These are user-specific cloud environments that are accessible through their own firewall. Private cloud environments are frequently set up as virtualized on-premises data centers. Users are able to use the virtualized infrastructure as a private cloud service thanks to an additional layer of automation.
Hybrid Cloud:
A hybrid cloud model is produced by combining public and private clouds. Because the workloads are movable across the hybrid cloud, businesses can use both the private and public clouds, depending on the workload’s security and economic considerations.