Web hosting provides access to remote server resources
Hosting is a contracted service that lets you use a provider's CPU, memory, storage, network and operational features to run websites, email or applications.
IBM describes web hosting as a provider supplying server space or a dedicated server for customers' websites and web applications. Offerings vary. In shared hosting, the provider commonly manages the OS and web server while customers configure their sites and email through a control panel. With VPS or cloud IaaS, customers may manage the guest OS, firewall, middleware, applications and data even when the provider manages facilities, physical hardware and virtualization. A dedicated server reserves physical resources but does not automatically include managed operations.
What the provider supplies
The contract determines which facilities, power, connectivity, hardware, virtualization, OS, control panel, monitoring, backups and support are included. The phrase “renting a server” alone does not define responsibilities.
What the customer manages
Customers often remain responsible for domains and DNS, accounts, websites, CMSs, plugins, applications, data, access rights and legal obligations. VPS and IaaS may also require customer-managed OS updates and network settings.
A server contract does not automatically build, protect or restore a website
Hosting is only part of the foundation. Separately review design, content, domains, DNS, TLS, CMS updates, access control, monitoring, backup scope and restoration procedures. Even with automatic backups, a recovery plan requires checking retention, frequency, coverage, restore granularity, cost and whether restoration actually works.
Choose WordPress hosting by its recommended requirements and operational features
“WordPress compatible” and storage capacity are not enough. As of November 3, 2026, WordPress.org recommends PHP 8.3 or later, MariaDB 10.11 or later or MySQL 8.0 or later, and HTTPS. Also consider updates, staging, backups and restoration, email, image processing, peak traffic and maintenance responsibilities together.
Official recommendations as of November 3, 2026
PHP 8.3+ / MariaDB 10.11+ or MySQL 8.0+ / HTTPS
Apache or Nginx is recommended. Older environments may still work, but unsupported versions can expose the site to vulnerabilities; check compatibility and update them.
Operations and recovery requirements
Operations = updates + staging + backups + restore tests + monitoring
Check automatic update scope, staging, backup coverage and retention, restore granularity, monitoring alerts and incident contacts. Test whether recovery meets your target time instead of relying on the presence of backups alone.
Example: a company WordPress site with tens of thousands of monthly views
- 1Inventory the workload: publish company information, inquiries and recruiting articles, maintained by two editors. There are no payments or member records, and email uses a separate service.
- 2Set requirements: recommended PHP, database and HTTPS support, staging, daily backups, activity logs, WAF coverage and support during Japanese business hours.
- 3Assign responsibilities: the provider manages the OS and infrastructure, the maintenance company handles WordPress core, themes, plugins and monitoring, and the in-house team manages editor access and publication approval.
- 4Test migration and recovery: verify URLs, forms, email, images, redirects and performance in a duplicate environment; restore from backup before switching DNS.
Managed shared hosting may fit these conditions, but reconsider VPS or IaaS if you need custom daemons or root access. Page views alone do not determine the right model.
Shared hosting, VPS, dedicated servers and cloud compared
Names differ by provider, so the table describes common configurations. IBM explains that shared hosting shares physical and software resources among sites; a VPS has its own OS and allocated resources on shared hardware; and dedicated hosting reserves one physical server. Cloud is distinguished by NIST characteristics such as on-demand access, resource pooling, rapid elasticity and measured service.
The table can be scrolled horizontally
| Service model | Resources and configuration | Suitable uses | Responsibilities and limits to check |
|---|---|---|---|
| Shared hosting | Customers share hardware, the OS and web and email infrastructure; settings are managed through a control panel. | Typical company websites, blogs and standard WordPress configurations. | Check root access, software versions, persistent processes, resource limits, effects of other tenants and backup coverage. |
| VPS | An isolated virtual server with allocated resources and its own guest OS on shared physical infrastructure. | Custom runtimes, APIs, test environments and workloads requiring OS or middleware control. | In self-managed service, the customer handles OS updates, SSH, firewall rules, monitoring, backups and incidents. |
| Dedicated server | One customer reserves the physical server's resources and may virtualize it into multiple environments. | Workloads requiring specific performance, configurations, licensing or physical isolation. | Exclusive use does not automatically improve availability. Check single-machine failures, spare hardware, replacement, redundancy and expansion time. |
| Cloud IaaS | Configure computing, storage and networking on demand through an API or control panel, with usage metered. | Variable workloads, multiple environments, automation and combinations of multiple zones or managed services. | Manage guest OS and application responsibilities, usage charges, network fees, access rights, configuration errors, limits and exit procedures. |
| Managed hosting | Adds operations such as OS updates, monitoring and backups to shared, VPS, dedicated or cloud hosting. | When you want to outsource infrastructure operations and focus on applications or business operations. | Managed describes an operational service, not a hardware type. Check covered operating systems and applications, support hours, recovery work and exclusions. |
| PaaS and static hosting | Abstracts runtime or delivery infrastructure so you deploy code or generated files without necessarily operating servers directly. | Static sites, web apps using supported runtimes and short development cycles. | Check supported runtimes, builds, persistence, execution time, portability and usage-based charges. These are alternatives to traditional hosting. |
VPS is not automatically better than shared hosting, nor dedicated better than VPS
More control usually brings more responsibility for configuration, updates, monitoring and recovery. Managed shared hosting with clear limits may be safer for a small company site. Consider VPS, dedicated or cloud services when you need root access, custom middleware, isolation or automation. Match the choice to your workload and operating capacity.
Six selection criteria to establish before signing up
Before comparing features, record current usage, peaks, acceptable limits and responsible people. Terms such as “fast,” “large” and “safe” are not decision criteria without measurement conditions and ownership.
1. Workloads and critical paths
Distinguish static company pages, WordPress, ecommerce, member sites, APIs, jobs and email. Record important paths such as inquiries, purchases and login, whether personal or payment data is involved, publication frequency and seasonal peaks.
2. Access rights and responsibilities
Assign root, SSH, SFTP and database access; updates to the OS, PHP, web server, CMS and plugins; vulnerability response, log review and certificate renewal among the provider, development company and in-house team.
3. Resources and performance
Check CPU, memory and storage, as well as limits on IOPS, concurrency, processes, database connections, execution time, transfer, bandwidth, files and email volume. Measure averages and peaks with the actual site.
4. Availability and recovery
Set acceptable downtime, RPO and RTO; check redundancy, monitoring, alerts, incident history and SLA scope. Review backup frequency, retention, location, encryption, coverage, restore granularity, request procedures and costs, and test restoration.
5. Security and support
Check MFA, access separation, WAF, DDoS and malware protection, audit logs, encryption, data location, deletion and incident notification. Distinguish support hours and channels from the scope of initial responses and recovery work.
6. Migration, exit and total cost
Include overages, network charges, backups, certificates, licenses, management features, support, migration, monitoring and staff time alongside setup and monthly fees. Check data export, DNS cutover, post-cancellation retention or deletion and lock-in.
Do not choose ecommerce or business-app hosting by traffic alone
The same number of views can produce different CPU, database, consistency and recovery needs for static pages versus search, inventory and payments. For ecommerce, plan for partial failures such as an order succeeding while notification or inventory updates fail. Check staging, monitoring, logs, idempotency, backups, recovery time and emergency contacts. A static information site may be simpler to run on static hosting without managing an always-on server.
Compare total hosting cost, not just the monthly price
A low monthly price can be followed by migration, backup restoration, monitoring, extra storage, network and maintenance costs. Compare total ownership costs over the same period and operating conditions.
Comparison formula
Total cost = setup + monthly fees + usage charges + migration + monitoring and maintenance + recovery + exit
Contract and usage fees
Check setup and monthly fees, minimum contract length, renewal prices, CPU, memory, storage, traffic, extra IP addresses and overages.
Related services
Separate charges for domains, DNS, email, certificates, CDN, WAF, malware protection, backup storage and restoration work.
Staff time
Include internal and external time for OS and CMS updates, monitoring, troubleshooting, inquiries, restore tests and billing.
Migration cost
Estimate data migration, test environments, DNS cutover, email migration, redirects, functional testing and running the old environment in parallel.
Incident and recovery cost
Allow for lost opportunities during downtime, emergency response, backup retrieval, recovery and prevention. Shorter RPO and RTO generally increase cost and complexity.
Exit and switching cost
Check data export, moving settings out of proprietary panels, retention or deletion after cancellation, domain transfer and lock-in.
Check eligibility and renewal terms for free SSL and domains
Benefits may depend on eligible domains, contract length, automatic renewal and cancellation terms. Compare total cost after promotional pricing ends and when switching providers.
Compare, contract and migrate in five stages
Choosing a provider before defining requirements can leave you without needed access or recovery methods. Measure the current system, test representative tasks and set acceptance criteria for cutover and exit.
- 1
Inventory the current configuration and usage
List domains, DNS, email, web services, CMSs, databases, certificates, external APIs, jobs, backups and owners. Measure average and peak storage, traffic, CPU, memory, response times, concurrency and incidents.
- 2
Define mandatory requirements and responsibilities
Separate required and desired OS, runtime and database versions, root access, data location, RPO, RTO, support hours and budget. Assign primary and backup owners for updates, monitoring, initial triage, recovery and user support.
- 3
Compare specifications, terms and SLAs consistently
Compare resource allocations, limits, maintenance downtime, backup, restoration, support, overages, data export and cancellation. Read the SLA's scope, measurement, exclusions, claim deadline and compensation separately from measured availability.
- 4
Run an acceptance test in a small real environment
Replicate representative pages, forms, login, search, image processing, email and jobs. Test peak load, logs, alerts, access rights, updates, backup and restoration. Compare your own configuration rather than advertised speed.
- 5
Plan migration, rollback and exit
Define data synchronization, DNS changes, TTL, certificates, email, freeze windows, acceptance, old-environment retention and rollback criteria. Continue reviewing billing, access, backups, monitoring and contacts after migration, and verify export and deletion at exit.
An SLA is neither a zero-downtime guarantee nor a recovery plan
An SLA defines contractual metrics and remedies for missed targets; it does not necessarily prevent every outage or data loss. Check exclusions for maintenance, customer configuration, external DNS and application failures. Prepare your own monitoring, redundancy, backups, restore tests and contact procedures.
Primary and official sources consulted
The following sources cover hosting definitions and types, cloud characteristics, IaaS responsibilities, WordPress recommendations and RPO/RTO. Product limits, prices, supported versions, SLAs, backups and operational scope can change; check the latest terms for each service before contracting.
Web hosting FAQ
How does shared hosting differ from a VPS?
Shared hosting generally shares the OS and web infrastructure among customers, with the provider managing those layers. A VPS has its own guest OS and allocated resources; in exchange for greater control, the customer may handle OS updates, SSH, firewall rules, monitoring and backups.
Do I need traditional hosting to use WordPress?
Traditional shared hosting is one option, not a requirement. As of November 3, 2026, WordPress.org recommends PHP 8.3 or later, MariaDB 10.11 or later or MySQL 8.0 or later, and HTTPS. Managed WordPress, VPS and cloud services can also run WordPress when their environments meet the applicable requirements.
Is a domain included with hosting?
A plan may include registration or renewal, but a domain and server have different roles and contracts. Record the registrant, renewal date, transfer conditions and DNS provider.
Does “unlimited” storage or transfer mean there are no limits?
Not necessarily. Fair-use terms and separate limits on CPU, memory, concurrency, files, databases, email, backups or bandwidth may apply.
Are automatic backups enough for disaster protection?
Not necessarily. Check scope, frequency, retention, location, encryption, restore granularity, request procedures and cost, and test actual restoration against your RPO and RTO.
Will a more expensive plan or faster CPU always speed up my site?
No. Measure the application, database, images, external APIs, cache and network to identify the bottleneck before adding resources or choosing improvements.
Should beginners choose shared hosting or a VPS?
For a typical company or WordPress site without an OS administrator, managed shared hosting or managed WordPress is a practical starting point. Consider VPS or IaaS for host OS root access, and compare VPS, PaaS or IaaS for custom runtimes and persistent processes after checking each service's limits.
Can I compare hosting by monthly price alone?
No. Compare total cost including setup, usage charges, backups and restoration, domains and email, monitoring, security, migration, maintenance time and data export at cancellation.
Explore related hosting terms
Choose hosting you can operate, not just the highest specifications
Hosting supplies resources and operational features to run websites and applications. Shared, VPS, dedicated and cloud services divide resources and responsibilities differently. Start with managed services for a standard company site; consider VPS or IaaS for root access and assess PaaS where its runtime and process limits fit. Compare workloads, supported software, peak performance, RPO, RTO, security, migration and total cost consistently, and test performance and recovery in a real environment.
How this article was reviewed
Finite Field Editorial Team
We consulted official IBM, NIST, AWS, WordPress.org and Google Cloud sources to distinguish definitions, service models, responsibilities, WordPress recommendations and recovery objectives. The article was checked July 28, 2026; WordPress recommendations and Cloud Run container privilege restrictions were rechecked November 3, 2026.