Great changes have been made since the turn of the 21st century, especially the revolutionary development by ICT converged technologies. It seems that we just stepped in the PC era yesterday and suddenly we are in the Cloud era today - an era of Cloud Computing and Coud services.
In the recent years, we have witnessed that Cloud Computing has spread across the world and directly changed the enterprise IT architecture and business models. So what actually are Cloud Computing and Cloud services? What key features do Cloud services have? How do they disruptively change the IT world? Follow up with me to find the answers to all these questions one by one.
CLOUD COMPUTING EVOLUTION
Source: https://www.toptut.com/2020/10/26/evolution-of-cloud-computing-its-future-prospects/
IT professionals may have noticed that enterprise IT infrastructure has become more complicated in the recent 10 years. This has translated as more problems ever than before. The typical problems include:
slow service rollout and long, yet complex product life cycle;
high Total Cost of Ownership (TCO);
key application performance limited by I/O bottlenecks.
To explain more, the traditional service rollout process consists of hardware procurement, software installation and deployment and service commissioning. This long process lasts several weeks to even months.
Additionally, with traditional IT architecture, a physical server often plays a single role, such as a SAP server, an Oracle database or a Web server. Therefore, specific software uses dedicate hardware, usage rate is high only during peak hours. Most of the time, the usage rate is nonetheless low and idle resources cannot be used by other services.
Unfortunately, it is harder to estimate the resource requirement and provide proper IT system capacity in IT architecture design. In fact, the provided capacity is either too much or too little. Too little sounds alright because we can expand the capacity, but what if it is too much? Return and refund? It is all but impossible. Capacity expansion is not that optimistic as you may think! This is the disadvantage of the traditional IT architecture where the expansion is difficult and lacks elasticity.
In a traditional computer system, the CPU I/O is faster than the memory I/O and the latter is faster than the disk I/O. This creates I/O bottlenecks when data are transmitted in the system: System component processing capability imbalance limits the application performance because of these I/O bottlenecks and limits. Customers can’t get the full return of their IT investment.
To sum it up, the traditional IT architecture is incapable to support the enterprise-level operation in an efficient manner. There is an urgent need for a new IT revolution. So what is this new revolution?
CLOUD COMPUTING DEFINITION BY NIST
Cloud Computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. This Cloud model is composed of five essential characteristics, three service models and four deployment models.
WHAT ARE CLOUD SERVICES THEN?
In short, Cloud services are a variety of Cloud Computing resources provided over the internet.
THE DEVELOPMENT OF CLOUD COMPUTING
Cloud Computing did not emerge all at once. It has gone through four distinct stages:
Virtualization -> Private clouds -> Data center consolidation -> Hybrid clouds.
With virtualization, physical computing, storage and network resources are grouped together into virtual pools. These resources are allocated to the VMs which do not rely on specific physical resources.
Virtualization decouples the OS and the software from the hardware, eliminating hardware compatibility issues once and for all. A hypervisor, or 'virtual machine monitor' manages the actual hardware and provides the VMs with hardware that is virtual.
Virtualization makes system maintenance (like a system migration or upgrade) much easier. For example, if the RAM or the CPU resources are not enough for your VM, the VM can be migrated to a more powerful system without even shutting down the VM. The VMs are not physical, they are logical, so they can be migrated from machine to machine easily, as long as the same storage and network are accessible.
When a virtual system gets too large, managing it manually can get expensive. Managing hundreds of hosts or thousands of VMs is far more complex and more expensive than smaller-scale deployments. This is why we need automated, process oriented, service-based cloud services.
Automated, process oriented, service-based cloud services
Source: https://www.appitel.fr/blog/appitel/les-differentes-solutions-de-cloud-computing-saas-paas-iaas/
Let’s take the Elastic Cloud Server, or ECS, as an example. As we saw earlier, users use the Web portal to specify the network, CPU and memory details they want for their ECS. Once their application is approved, the ECS is created for them automatically in the Cloud. When the ECS is no longer needed, the user can just delete it!
This self-service process is the life-cycle of an ECS: apply, approve, build and release. The entire lifecycle of a VM can be automated. Not only that, VM resources can be scheduled elastically. For example, a user who applied for just 10 ECSs originally can add 10 more if their demand suddenly increases or reduce drop to 5 to meet if their service requirements decrease.
The Huawei public Cloud also supports high availability, or HA and distributed resource scheduling, or DRS. We will discuss these features in the next article.
THE FEATURES OF CLOUD COMPUTING
The industry claims that Cloud Computing is an on-demand self-service with various network access compatibilities. It pools computing, storage and network resources. This allows it to flexibly scale up and down based on the requirements, while charging according to time, traffic, or performance.
Source: https://codete.com/blog/what-is-cloud-computing-software-development-benefits-and-challenges-for-business-owners/
You can use the web UI to apply for various Cloud services in accordance with your resource requirements. For example, ECS allows you to apply for a VM online. You can specify the number of CPUs, memory size, storage capacity and I/O performance parameters before pressing ‘Submit’. Then your required ECS will be created in the Cloud backend. The convenience of this process is self-evident. You are allowed to access ECS from home or at work using the internet or VPN, respectively.
Huawei Public Cloud - ECS Creation
Huawei Public Cloud - ECS Creation
All the resources used by ECS are logical resources that can be scaled up and down at any time, according to your demand. For example, if you are just starting a business, you only need to purchase a few ECS resources at first. You can add resources using Auto Scaling as your service requirements grow.
During your effort to cut costs, you are allowed to decrease the ECS resources during off-peak hours. This on-demand resource allocation model saves you money.
In this article, we have learned about the background, definition, the development and basic features of Cloud Computing. In the next article, I will introduce Huawei’s perspectives on Cloud Computing and Cloud service.
