Addendum to the recommendations of the IT Infrastructure Commission motivation
The use of cloud services is a significant IT trend. For infrastructure operations, moving computing and storage services to cloud service providers has become an interesting alternative that promises flexibility
and cost-efficiency benefits. The trend is supported by the private sphere, where users regularly access cloud-based services from mobile and non-mobile users. Commercial cloud providers provide numerous attractive services and tools for this purpose.
In the field of science, this raises the question as to what extent cloud-based offerings represent an alternative or supplement to the previous infrastructure solutions. This raises questions about the funding and financing possibilities of such services, as it is a transition from traditional investment and the operation of equipment and systems to a billing of services.
The use of cloud services brings with it advantages as well as risks in terms of security, privacy and sustainability. Therefore, a closer look at these offers within the framework of the scientific IT infrastructure is necessary. In the following, some clues will be given.
Definition of cloud services
The term cloud computing as a generic term for cloud-based infrastructures, services and applications is used in different contexts. The most commonly mentioned definition comes from the National Institutes
of Standards and Technology (NIST) and reads as follows:
Cloud computing is a model for enabling, convenient, on-demand network access to a shared pool of configurable computing resources (eg, networks, servers, storage, applications, and services) service provider interaction. This cloud model is composed of five essential characteristics, three service
models, and four deployment models.
According to this definition, cloud-based infrastructures are characterized by the following properties:
- The provision of resources is highly automated and without human interaction (on-demand, self- service).
- Network-based access to resources is provided through standardized interfaces.
- Existing resources are dynamically distributed among requesting users, regardless of the geographic location of resources and users.
- Resources can be booked or released at short notice (elasticity).
- The resource usage is logged and billed. The main service models are (Iaas, PaaS, SaaS):
- Infrastmcture as a Service (IaaS): the operator provides the user with an infrastructure such as virtual machines or storage. The user usually has no influence on the physical infrastructure.
- Platform as a Service (PaaS): the user can implement their own applications on predefined platforms (libraries, databases, computing, analysis solutions) and does not have to operate them themselves.
- Software as a Service (SaaS): in this case, the operator provides a complete application. The user has no control and, if necessary, knowledge of the underlying infrastructure or the place of operation.
An essential differentiator is the operating model. This is between
- Hybrid or
- Community Clouds distinguished.
A public cloud operates on external resources of a provider and can be accessed by a broad public, while the private cloud is available to an a-priori defined group of users (e.g., members of an organization) mostly on local resources. Besides, there are hybrid forms in which services e.g. within cross-community user groups (community clouds), or local ones connected to external resources (hybrid clouds).
It is foreseeable that there is a growing demand for cloud-based infrastructures. The benefits are very diverse for both cloud users and cloud providers. Advantages for the users are the saving of acquisition costs for hardware and software as well as the possibility of an individual scaling of the required capacities as needed and without a longer binding, as
it results from a procurement of hardware. Furthermore, the effort for maintenance, further developments, license and personnel costs can be reduced to a minimum. Shifting operational responsibility to the cloud provider also reduces time and costs.
Another important advantage is the potential use of common platforms for different facilities and user groups. Thus, work on joint projects in the field of research as well as in teaching can be considerably simplified in the case of a decentralized distribution of users through the use of shared storage, arithmetic or library clouds. Users of cloud offerings can test and develop new concepts with low initial costs. Cloud-based infrastructures also offer many advantages for the cloud provider. You can distribute the hardware and software acquisition costs to more customers, highly standardize and automate them accordingly. The associated “economy of scale” reduces administrative and operational costs, for example by increasing the utilization of individual systems and optimizing energy costs.