Bare metal – the latest product on the web hosting market evokes power from its name alone, and not for nothing: servers with this nickname focus on high-performance hardware. With a bare metal server, the user's access extends practically down to the bare metal. Software components based on this product are subject to user control and can be installed and configured according to individual...
Setting up your own data center is time consuming and expensive: there are considerable costs involved for maintenance and for hiring specialist staff. In addition, it is difficult to scale as needed. In order for your business to grow, your IT infrastructure must always be one step ahead, and therefore you often need to provide more resources than your stuff currently need. A fully functional data center is vital for most companies, and work would stand still if it were to fail. Nobody can afford that, so it’s better to invest too much in a data center rather than too little.
With cloud computing, you can go another way: instead of buying, installing, and maintaining the technology yourself, you use web services. Various services are available from a data center via the internet, including storage space, processing power, databases, and software environments. In most cases, this form of outsourcing IT resources is run on a rental basis. In this way, services are easy to book precisely when you need them.
History of cloud computing
Networking computers to provide users with more processing power and storage space is not a new invention. Computer mainframes in the 1950s had already taken a step towards cloud computing. At that time users could access the mainframe computer via several other terminals in the organization and use their capacities. However, this was on a timeshare basis to begin with, where users had to reserve times to use the computational power of the mainframe.
In the following decades virtualization was developed, which allowed computer instances to be recreated abstractly. Finally, with the invention of the internet, such virtual environments were available to everyone online and became commercially available to greater numbers of users in the 1990s.
At this point, the concept of the cloud became more popular, but it wasn’t until the millennium that companies and private individuals began to become increasingly interested in technology. The first cloud offers were individual services, such as a space to share files or Google Sheets and Docs, where multiple users can work simultaneously on a document. However, at the same time, Amazon also began to provide its enormous server farms to other users. Known as Amazon Web Services (AWS), it allows other companies to use the infrastructure of the e-commerce giant and run software on it.
Nowadays, cloud computing is part of everyday life for many people. Most smartphones – the internet of things even more so – are constantly connected to the cloud. For example, photos that users take with their smartphone are uploaded automatically to Apple or Google’s cloud service so they can be accessed from other devices.
What is cloud computing?
Cloud computing is a collective term describing hardware and software services provided via the internet. The extent to the services offered can vary from cloud storage space facilitated by remote servers to infrastructure in the cloud, with which companies gain access to complete data centers over the internet.
The collective term cloud computing describes the supply of various hardware and software solutions via the internet. Users can rent processing power, storage space, and software environments to upgrade or expand their existing infrastructure.
According to a paper by the National Institute of Standards and Technology, services must fulfill certain characteristics in order to be called cloud computing:
- On-demand self service: It should always be possible for users to independently request the resources they need without having to contact the service provider.
- Broad network access: Cloud computing must be accessible over the internet and through standardized mechanisms and protocols that ensure widespread accessibility.
- Resource pooling: Pooling multiple computing resources is a basic requirement for cloud computing. This is done in the form of server farms, which dynamically assign and reassign resources, such as processing power and storage, to satisfy the demand of multiple users. Customers may not know the precise location of the resources provided.
- Rapid elasticity: The delivery of capacities must be fast and needs based. In some cases, this may be automated to scale up and down without the intervention of customers or service providers.
- Measured service: Cloud service use is monitored at all times, creating greater transparency for both parties involved.
Cloud computing is often understood as an attempt to make IT resources available in a similar way to other infrastructural services: just as water comes from a tap and electricity comes from a socket, computing power comes from the internet. In the same way that most people don’t have to understand how electricity networks function, cloud computing eliminates the need to understand how a computer is composed. The user simply uses the resources (storage space and computing power) and receives a bill for it, exactly like electricity usage.
Like the power station is the center of the power grid, cloud computing is usually also centered around a large data center or a server farm. This is where the resources of multiple computers or servers are pooled together, known as grid computing, which enables high performance. In combination with virtualization, individual virtual instances can be created for individual users within the network. The network access works seamlessly, so the user doesn’t have to know precisely where their data is stored.
Varieties of cloud computing
There are now a wide array of cloud solutions on the market. Besides the price and support provided, these options differ mainly in the layers on offer and the deployment model. The term “layers” refers to the scope of the service and the delivery model identifies the type of offer.
A layer corresponds to a specific service level, and the various “as a service” levels, or layers, describe the scope of the service. For this reason, they are usually presented using a pyramid model. Infrastructure as a service has the largest scope, while software as a service focuses only on specific applications.
- Infrastructure as a service (IaaS): On this level, providers offer complete hardware solutions: processing power, storage space, and network technology. The instances the user utilizes are entirely virtual and are divided into the resource pool. IaaS can serve as a basis for further layers, but is also offered as a standalone product.
- Platform as a service (PaaS): This layer goes a step further by providing a complete cloud environment rather than just the hardware. PaaS is primarily targeted at software developers. Service providers provide an established development environment in the cloud on hosted hardware. Programmers therefore save time and money on setting up and maintaining such an environment.
- Software as a service (SaaS): At the highest layer, users are offered software solutions from the cloud. SaaS is mainly intended for a typical end user, since they don’t have to worry about installing and maintaining software, and have the reassurance that the hardware is sufficiently powerful. To access the software, users either use a web browser or a limited program that loads the software from the cloud.
- Everything as a service (XaaS): As well as the three layers mentioned above, providers also offer further services. However, this is normally for marketing reasons. A XaaS can in fact always be attributed to a different level, or have nothing to do with cloud computing. Humans as a service (HuaaS) is a form of crowdsourcing where a pool of people undertake work over the internet.
The deployment models describe the type of service, such as whether the instances are reserved for just one user or shared with others.
- Private cloud: The servers are used exclusively by one customer. A private cloud can be on-site (internal cloud), but doesn’t have to be. Even with a hosting provider who uses a server farm, it’s possible to use dedicated hardware which other customers of that provider cannot access.
- Community cloud: The community cloud functions similarly to a private cloud, except that multiple users share a dedicated hardware instance. However, the pooling of users is not done at random. Instead, customers from the same business fields or with similar interests are often grouped together. Furthermore, the community cloud can either be managed in a company or externally. The objective is to save resources compared to running multiple private clouds.
- Public cloud: This type of deployment model corresponds to the actual idea behind the cloud. That is, you share a server connection with the general public. As a user, you cannot view or alter who else is using the server resources.
Hybrid cloud: This model is a hybrid of the private and public clouds. This means commercial and private users can determine which portions of their operation to make private – security aspects for example – and which to leave public.
Advantages and disadvantages of cloud computing
Cloud computing has significant advantages, especially for small to medium-sized organizations. For such enterprises, establishing their own IT infrastructure is costly and requires continuous support and maintenance from dedicated staff. In addition, it’s best for a data center to grow proportionally with the growth of the organization itself. Since this is traditionally difficult to achieve, decision makers tend to buy new hardware in set cycles based on an estimate of their future needs. This often leads to more resources being purchased than are actually needed.
Nevertheless, most companies cannot operate without a data center. Cloud computing provides a needs-based solution to this issue. The acquisition and maintenance of the hardware is taken over by the service provider, and the customer need only invest relatively cheap terminals in order to access the cloud computing services. In addition, professional server farms, such as those used by cloud providers, are much more secure than most organizations could manage on-site. Dedicated security personnel and server specialists protect the data center from physical and digital attacks respectively, and fire safety experts ensure data is safe from fire. Finally, most cloud computing service providers conduct compulsory backups of all data.
However, cloud computing also has some drawbacks, which is why some have still not made the switch. For example, the level of customizability in terms of configuration can vary greatly depending on the service provider, which is a significant disadvantage for some. Generally speaking, you are also dependent on the provider – if they are having technical difficulties, it will directly affect your operations. Furthermore, cloud computing requires a strong and stable internet connection to be effective. If this isn’t in place, employees won’t be able to work efficiently.
The largest concern over cloud solution is about data privacy. Although it’s true that data is secured in the data center or server farm, transmitting that data over the internet always presents a security risk. Furthermore, the location of the data center also has an effect. While US-based providers should reasonably be expected to adhere to local data protection regulations, the same cannot necessarily be said of international providers.
Of course, although cloud computing can help companies save money, it’s not free, so users must think carefully about how much capacity they need and when. Companies often book more cloud resources than they actually because of the relatively low cost at face value, but these costs add up. This doesn’t have to be the case, however, since some providers allow you to cancel cloud instances as soon as they are no longer needed, saving you money.
No acquisition costs
Requires fast and stable internet connection
No capital commitment
Data privacy concerns
Scalable as required
Dependency on the service provider
No in-house specialist required
Security risk during data transmission over the internet
Data centers are well maintained and secure
Low individual prices make it tempting to book more resources than you need