Cloud Computing Terms: IaaS, Virtualization, VPS and Storage Glossary

Block storage keeps data in fixed-size blocks and presents it to the operating system as a raw disk (volume). The file system is created by the server that mounts the disk. Virtual machine disks, SAN systems and database volumes are typical examples.

What it means for a business: It is used for database and operating system disks that need low latency and high IOPS; it costs more per gigabyte than object storage, so archive data usually moves to object storage.

Related: IOPS · object storage

Cloud computing is a model for on-demand network access to a shared pool of configurable computing resources such as servers, storage, networks and applications. NIST SP 800-145 lists five essential characteristics: on-demand self-service, broad network access, resource pooling, rapid elasticity and measured service. Resources are provisioned and released with minimal interaction with the provider.

What it means for a business: Capacity is rented as needed instead of bought up front, turning capital expenditure into operating expense and adding capacity in hours rather than weeks.

Related: NIST SP 800-145 · cloud solutions · IaaS

Cloud migration is moving applications, servers and data from on-premises infrastructure or another provider into a cloud environment. Common approaches are moving without changes (rehost, lift-and-shift), partial adaptation (replatform) and redesigning the application (refactor). Inventory, dependency analysis, testing and a rollback plan are the core steps.

What it means for a business: Downtime and data consistency are the biggest risks; measuring readiness before the move makes the plan much clearer.

Related: cloud migration services · migrating servers to the cloud · cloud migration readiness assessment

A cloud server is a virtual server running on a cluster of physical hosts with shared storage. If one host fails, the virtual server can be restarted on another host in the cluster, and its CPU, memory and disk can be resized as needed. In practice it is the most common IaaS product.

What it means for a business: It offers higher availability than a VPS tied to a single machine and is the usual starting point for ERP, web applications and file servers.

Related: cloud server rental · benefits of cloud servers · VPS

A container packages an application with its dependencies and runs it as an isolated process that shares the host operating system kernel. Unlike a virtual machine it carries no kernel of its own, so it is lighter and starts in seconds. Docker and containerd are widely used container runtimes.

What it means for a business: Containers make an application behave the same in every environment; they usually run on virtual machines or physical servers, so they sit on top of virtualization rather than replacing it.

Related: Kubernetes · virtual machine · Linux server management

Data residency is the country or region where data is physically stored and processed; the term also covers requirements, set by law, contract or internal policy, to keep data within a given geography. Data sovereignty is a related but separate idea: that data is subject to the laws of the country where it resides.

What it means for a business: Under Türkiye’s KVKK, transferring personal data abroad is subject to extra conditions, so keeping data in Türkiye simplifies compliance.

Related: KVKK and data backup · KVKK

A dedicated server is a physical server rented in full to a single customer. There is no provider-side virtualization layer; CPU, memory and disks are not shared with anyone else, and the customer may install its own hypervisor. The term bare-metal server is used with the same meaning.

What it means for a business: It suits software licensed per physical core, databases that need high disk performance and workloads that require full hardware control; plan separately for what happens if the hardware fails.

Related: dedicated server rental · colocation · cloud server

ESXi is VMware’s Type 1 hypervisor, installed directly on a physical server. It needs no separate general-purpose operating system and shares CPU, memory, storage and network resources among virtual machines. A single host can be managed on its own, but multiple ESXi hosts are usually managed centrally through vCenter.

What it means for a business: ESXi versions and patches matter for security, because a hypervisor vulnerability can affect every virtual machine on the host.

Related: hypervisor · vCenter · VMware management

A GPU server is a server that has one or more graphics processing units (GPUs) alongside its general-purpose CPU and is built for workloads that need parallel computation. AI model training and inference, image processing, 3D rendering and scientific computing are typical uses. GPU memory (VRAM) is the main limit on which models can run.

What it means for a business: For periodic heavy computation, renting instead of buying reduces the risk of expensive hardware becoming outdated quickly.

Related: GPU server · what is a GPU server · GPU VRAM calculator

Horizontal scaling means adding servers that do the same job (scale out) or removing them (scale in) to match demand. Incoming traffic is spread across the servers by a load balancer. The application must be designed to run as several copies and to keep session state in a shared place.

What it means for a business: It removes the single-server ceiling and keeps the service running if one server fails; it suits web applications and RDS session hosts.

Related: load balancer · vertical scaling · RDS server sizing calculator

A hybrid cloud is a composition of two or more distinct cloud infrastructures (private, community or public) that remain separate but are bound together by technology that enables data and application portability; this is the NIST SP 800-145 definition. In practice the term is also used for on-premises servers working together with a cloud environment.

What it means for a business: Sensitive or hard-to-move systems stay in place while backup, disaster recovery or seasonal load moves to the cloud; networking and identity between the two are central to the design.

Related: cloud solutions · multi-cloud

A hypervisor is the virtualization component that runs between physical hardware and virtual machines and manages the guest operating systems; NIST also notes that it controls how guests communicate with the hardware. A Type 1 (bare-metal) hypervisor runs directly on the hardware, while a Type 2 hypervisor runs on top of an operating system. VMware ESXi and Microsoft Hyper-V are Type 1 examples.

What it means for a business: Servers use Type 1 hypervisors, and the choice directly affects licensing, management tools and backup software compatibility.

Related: NIST glossary: hypervisor · ESXi · virtualization

IaaS is the cloud service model in which basic infrastructure resources such as processing, memory, storage and networking are delivered as a service. The customer manages the operating system, applications and data; the physical hardware and virtualization layer are the provider’s responsibility. Renting cloud servers and virtual machines is the most common example.

What it means for a business: It is the easiest way to move existing applications unchanged; OS patching and backups stay with you unless handed over as a managed service.

Related: cloud server rental · PaaS · managed services

IOPS (input/output operations per second) is the number of read and write operations a storage system can complete each second. The figure depends on block size, read/write mix, random versus sequential access and queue depth, so an IOPS number quoted without those conditions cannot be compared. It is read together with latency and throughput.

What it means for a business: Slow database and RDS servers are often limited by disk rather than CPU, so ask about disk type and IOPS limits before signing.

Related: block storage · RDS server sizing calculator

Kubernetes is a portable, extensible, open-source platform for managing containerized workloads and services. It automates placing containers on servers, scaling them, restarting them after failures and rolling out updates, driven by declarative configuration. The project is hosted by the Cloud Native Computing Foundation (CNCF).

What it means for a business: It is a strong tool for teams running many microservices, but it takes real expertise to run; for a handful of applications, virtual machines are often simpler.

Related: Kubernetes documentation · container

A load balancer is a network component that distributes incoming requests across several servers according to an algorithm. Health checks remove servers that stop responding. It can work at the transport layer (L4, using IP and port) or the application layer (L7, using HTTP content) and comes as hardware, software or a cloud service.

What it means for a business: It is used both to add capacity and to keep the service up during maintenance, and is often paired with TLS termination and a web application firewall.

Related: horizontal scaling · WAF

Managed services means handing the monitoring, maintenance and troubleshooting of IT functions such as servers, networks, backup or security to a service provider under contract. Scope, response times and the limits of responsibility are usually defined in a service level agreement (SLA).

What it means for a business: Organizations with small IT teams can hand over routine work such as 24/7 monitoring and patching; the contract should state clearly which tasks stay in-house.

Related: managed services · service level agreement

Multi-cloud is the use of services from more than one cloud provider at the same time. Its focus is on multiple providers rather than on mixing deployment models, although the two terms overlap and there is no industry-wide agreement on a strict boundary with hybrid cloud. Typical goals are reducing dependence on one vendor and picking the best-fitting service for each workload.

What it means for a business: It adds flexibility but multiplies monitoring, security policy, billing and skills requirements by the number of providers.

Related: hybrid cloud · managed services

Object storage is a storage architecture that keeps data as objects, each made of a unique key, the data itself and metadata, instead of a file hierarchy or disk blocks. Objects live in containers called buckets and are accessed through an HTTP-based API. Capacity grows horizontally by adding nodes.

What it means for a business: It is a cost-effective tier for backups, archives, media files and logs, but it is not designed for low-latency uses such as database disks.

Related: S3 object storage · S3-compatible storage · block storage

Overcommitment is allocating more total vCPU or memory to virtual machines than the physical host actually has. It relies on the assumption that not all virtual machines run at full load at the same time. For example, assigning 96 vCPUs on a host with 32 hardware threads is a 3:1 overcommit.

What it means for a business: Moderate ratios improve hardware efficiency; at high ratios virtual machines wait in the CPU queue (CPU ready) and performance drops at peak hours.

Related: vCPU · VDS · choosing a VPS

PaaS is the cloud service model in which the platform needed to build and run applications (runtime, database, middleware) is delivered as a service. The customer manages only the application and its data; the operating system, servers and network are run by the provider. NIST SP 800-145 defines PaaS as one of three service models alongside IaaS and SaaS.

What it means for a business: Developers can focus on the application instead of servers, within the languages, versions and settings the platform supports.

Related: NIST service models · IaaS · SaaS

A private cloud is a deployment model in which the infrastructure is provisioned for the exclusive use of a single organization. Under the NIST definition it may sit on the organization’s premises or in a provider’s data center, and it may be operated by the organization, a third party or both. Resources are not shared with other customers.

What it means for a business: It suits workloads that need strong isolation, predictable performance or specific compliance controls; it usually costs more than shared infrastructure.

Related: cloud solutions · public cloud · hybrid cloud

A public cloud is a deployment model in which a provider operates the infrastructure and serves many customers from a shared resource pool. Customers share the same physical hardware, separated by virtualization and access controls. Resources are obtained by self-service over the internet and billed by usage.

What it means for a business: It offers a fast start and wide scaling options; check in advance which country the data is stored in and how the monthly bill changes with usage.

Related: benefits of cloud servers · data residency · private cloud

RDS (Remote Desktop Services) is the Windows Server infrastructure that lets authorized users connect over the network to a server-hosted desktop or to individual applications (RemoteApp). Applications run on the central host and only the user interface is sent to the client over RDP. Formerly called Terminal Services, it is unrelated to Amazon RDS, a database service.

What it means for a business: A common way to run ERP and accounting software for branches and remote staff; each user or device needs an RDS CAL.

Related: Microsoft RDS overview · RDS server sizing calculator · Windows Server management

S3-compatible storage is an object storage service that supports the Amazon S3 API (REST operations such as PUT, GET and LIST, plus its request-signing method) and therefore works with applications written for that API. The level of compatibility varies by provider, so support for features such as Object Lock, versioning and lifecycle rules should be checked separately.

What it means for a business: Most backup software can target S3-compatible storage directly, so data can be written to storage located in Türkiye without changing the software.

Related: Amazon S3 API Reference · S3 object storage · Object Lock

SaaS is the cloud service model in which an application runs on the provider’s infrastructure and is delivered to users by subscription, usually through a browser or client app. The provider manages the application, infrastructure and updates; the customer controls its own data and a limited set of application settings. Email, office suites and CRM software are typical examples.

What it means for a business: Installation and maintenance disappear, but protecting the data remains the customer’s job, which is why Microsoft 365 data is often backed up separately.

Related: Microsoft 365 backup · PaaS

vCenter Server is the VMware management server that controls multiple ESXi hosts and their virtual machines from a single interface. Clustering, live migration with vMotion, high availability, Distributed Resource Scheduler (DRS) and permissions all run through vCenter. It is usually deployed as a prebuilt virtual appliance.

What it means for a business: Access to vCenter means access to the whole virtual estate, so administrator accounts should be protected with multi-factor authentication and least privilege.

Related: VMware vSphere · multi-factor authentication

A vCPU (virtual CPU) is a processor unit assigned to a virtual machine and scheduled by the hypervisor onto physical cores or hardware threads. One vCPU usually corresponds to one hardware thread, but the mapping and the overcommit ratio vary between providers.

What it means for a business: The same vCPU count can perform differently at different providers, so include CPU generation and overcommit ratio when comparing offers.

Related: RDS server sizing calculator · overcommitment

A VDS is usually described as a virtual server whose CPU and memory are reserved for it and not shared with other customers. The term has no standard definition: some providers use VDS and VPS interchangeably, while others use VDS to signal that resources are not overcommitted.

What it means for a business: When comparing offers, look past the VDS label and ask whether vCPUs are pinned to physical cores and what overcommit ratio applies.

Related: VDS vs VPS vs cloud server · overcommitment

Vertical scaling means adding CPU, memory or disk to an existing server (scale up) or removing them (scale down). It requires no change to the application, but it eventually hits the limits of a single server and its underlying host, and some changes need a reboot.

What it means for a business: For databases and ERP systems designed to run on one server it is the most practical way to grow; if a reboot is needed, the change is scheduled in a maintenance window.

Related: horizontal scaling · cloud server rental

A virtual machine is a software-based computer with its own operating system that runs on CPU, memory, disk and network resources allocated by a hypervisor. It is isolated from other virtual machines on the same physical server. Because it is stored as a set of files, it can be copied, moved and backed up as a whole.

What it means for a business: Running each application in its own VM keeps failures and updates from affecting each other, and the VM can be resized as needs change.

Related: cloud server rental · vCPU · container

Virtualization is the use of a software layer to divide one physical server’s CPU, memory, storage and network resources among several isolated virtual machines. That layer is provided by a hypervisor. Storage, network and desktop virtualization apply the same principle.

What it means for a business: It raises hardware utilization, cuts server provisioning to minutes and enables operational tools such as snapshots and live migration; it is the technical foundation of cloud server services.

Related: what is VMware · VMware management · hypervisor

VMware vSphere is the name of VMware’s server virtualization platform, built mainly from the ESXi hypervisor and the vCenter management server. Features such as virtual machine provisioning, resource pools, high availability (vSphere HA) and live migration of running virtual machines (vMotion) are delivered on this platform. VMware has been part of Broadcom since 2023.

What it means for a business: It is a common standard in enterprise virtualization; after the licensing model changes, many organizations are comparing the cost of owned versus rented infrastructure again.

Related: VMware management · what is VMware · vCenter

A VPS is a virtual server carved out of a physical server, with its own operating system and administrator access. Resources are shared with other VPSs on the same host and are often overcommitted. In a classic VPS setup the virtual server is tied to one physical machine, so a failure of that machine stops the VPS too.

What it means for a business: It fits low-cost websites and test environments; for critical applications, ask how much capacity is reserved and whether the hardware is redundant.

Related: how to choose a VPS · VDS · cloud server