Shared, Reseller, VPS, Dedicated, and Colocation Hosting

Hosting products are often compared by CPU, memory, disk, and bandwidth. Those specifications matter, but the more important difference is where responsibility changes hands.

The Responsibility Ladder

Shared hostingProvider manages nearly the entire platform.

Reseller hostingProvider manages the platform; reseller manages customers and packages.

Managed VPSProvider manages virtualization and an agreed operating-system/application scope.

Unmanaged VPSProvider manages the virtualization platform; customer manages the guest.

Dedicated serverProvider manages facility and hardware; operating-system responsibility depends on the contract.

ColocationProvider supplies space, power, cooling, and network; customer owns the equipment.

Side-by-Side Comparison

Model	Isolation	Provider responsibility	Customer responsibility	Best fit
Shared	Account-level	Server, control panel, web stack, mail stack, backups, patching	Application, content, account security	Small sites and customers who do not want server administration
Reseller	Multiple accounts under a reseller boundary	Same platform responsibilities as shared hosting	Packages, customer support, billing, account-level configuration	Designers, agencies, and small IT providers
Managed VPS	Virtual machine or container	Hypervisor plus contracted management scope	Application ownership and approved changes	Workloads needing isolation without a full internal operations team
Unmanaged VPS	Virtual machine or container	Host, storage, network, and virtualization layer	Guest OS, firewall, updates, services, backups unless purchased separately	Administrators and developers comfortable managing Linux or Windows
Dedicated	Physical server	Facility, network, power, and hardware replacement	OS and applications unless managed service is included	Consistent performance, licensing needs, large storage, or special hardware
Colocation	Customer-owned hardware and allocated rack space	Space, power, cooling, physical security, network handoff	Hardware, spares, OS, application, lifecycle, and often remote-hands requests	Organizations wanting hardware ownership without operating a facility

Shared Hosting

Shared hosting is an exercise in controlled multi-tenancy. Many accounts share the same operating system, web server, database service, and mail system. The product succeeds when accounts are sufficiently isolated, noisy neighbors are contained, and routine work is automated.

What the operator must control

Per-account resource use, including CPU, memory, process count, I/O, and entry processes
PHP versions and application compatibility
Mail reputation and outbound rate limits
File ownership and account isolation
Backup schedules and restore access
Malware, outdated CMS installations, and compromised credentials

Real-world example: A single compromised WordPress site can create a mail-reputation problem for every customer sharing the same outbound IP. Shared hosting therefore requires application-level abuse controls even when the provider did not build the application.

Reseller Hosting

Reseller hosting adds a business layer on top of shared hosting. The reseller may create accounts and packages, but the underlying platform remains the provider’s responsibility.

The contract should clearly answer:

Does the reseller provide first-line support to end users?
Who handles abuse complaints and legal requests?
Can the reseller oversell disk or bandwidth?
Who owns backups and restores?
What happens to end-user accounts if the reseller stops paying?

Virtual Private Servers

VPS hosting provides a stronger technical boundary, but “managed” and “unmanaged” are commercial terms rather than universal technical definitions. Two providers may sell managed VPS service with very different scopes.

A useful managed-service matrix

Task	Included?	Response target	Notes
Operating-system security updates	Define explicitly	Routine or emergency	Include reboot expectations.
Control-panel administration	Define explicitly	Business hours or 24/7	Clarify supported panels.
Application debugging	Usually limited	Best effort	Provider may prove the platform is healthy without fixing custom code.
Backup monitoring	Separate from backup storage	Daily review	A backup product is incomplete if failures are not acted upon.
Incident cleanup	Often billable	Emergency	Define whether forensic work is included.

Dedicated Servers

A dedicated server removes the hypervisor as a shared performance layer, but it introduces hardware lifecycle decisions. The operator must know what is stocked, what can be replaced quickly, and what requires a full system migration.

Questions beyond the processor model

Are the boot disks mirrored?
Is the RAID controller battery or cache module monitored?
Is out-of-band management included and isolated?
How quickly can memory, drives, power supplies, and system boards be replaced?
Is the server a standard build or a one-off configuration?
What is the migration plan when the hardware reaches end of life?

Colocation

Colocation moves the provider’s responsibility down to the facility layer. That may look simpler, but power density, cabling, access control, remote hands, and network demarcation become the product.

The colo demarcation must be visible

Provider responsibility                    Customer responsibility
-----------------------                    -----------------------
Rack / cabinet                             Server and storage hardware
Power circuits and metering                Power supplies and power cords after handoff
Cooling and facility environment           Airflow within equipment
Network cross-connect / switch port        Customer router, firewall, or server NIC
Physical access control                     Equipment configuration and data
Optional remote hands                       Written instructions and replacement parts

A colocation customer should know whether the network handoff is an access VLAN, a tagged trunk, a routed subnet, or a point-to-point circuit. They should also know whether remote hands can swap a drive, attach a crash cart, move a cable, or only perform visual checks.

Choosing the Right Model for a Customer

Start with operational skill. A customer who cannot maintain an operating system should not be sold an unmanaged VPS merely because it is inexpensive.
Identify isolation requirements. Compliance, performance, software dependencies, and reputation may require a separate VM or server.
Understand growth pattern. Predictable steady growth and bursty workloads may benefit from different platforms.
Match recovery expectations. A customer expecting near-immediate recovery needs more than a nightly backup.
Price the support burden. The least expensive infrastructure may become the most expensive product to support.

Common Product Design Mistakes

Unlimited language

Every system has limits. Replace vague “unlimited” promises with fair-use boundaries and measurable resource policies.

Undefined management

“Managed” without a task list creates a different expectation for every customer.

Backups sold as magic

State frequency, retention, exclusions, restore process, and whether customers should maintain independent copies.

One-off hardware

Special configurations can be profitable, but they need spare-parts and replacement plans.