Cloud Computing IaaS versus Dedicated Server Hosting You have a number of infrastructure hosting options when it comes to the type of server, storage and network architecture you need for your applications, databases and websites. Some Cloud computing servers now perform and behave exactly like a dedicated server, providing greater reliability, flexibility, scalability, and incredible performance, while giving unprecedented flexibility, and elasticity while costing much less. Below is a side-to- side comparison of the two most popular options to help you select the type that is best for your application needs.
| Cloud Servers | Dedicated Servers | |
|---|---|---|
| Definition | Simply put, a true Cloud server or instance is hardware that operates via virtualization and the provider's provisioning software, allowing you to modify key resources. Need more CPU power? Need more memory? Storage? All physical server resources can be adjusted on-demand without having to wait for a technician to power-down the server, open the server, install new RAM, and repower. | A dedicated server is a type of Internet hosting in which the client leases an entire server not shared with anyone else. This is more flexible than VPS hosting, as organizations have full control over the server(s), including choice of operating system, hardware, etc. |
| Operating System | Dedicated Operating System; Support any x86 compatible operating system (Unix, Linux, MS Windows, and more) | Dedicated Operating System; Most dedicated hosting services require you to select from a list of pre-approved OS and versions. |
| Software Applications | Customer can remotely install and manage any compatible software application | Customer can remotely install and manage any compatible software application |
| Administrative Access | Full Root access for Unix/Linux-based Servers. Full Administrative access for Windows Servers | Root access for Unix Servers. Administrative access for Windows Servers |
| Remote OS Install/Reinst all | Yes | Rarely |
| Remote Server Console Access | Yes | Yes |
| Remotely Reboot a “frozen” Server | Yes | Sometimes |
| Server Snapshots | Customer can take a server snapshot before making changes to it. A simple mouse click will return server back to the original state if needed. | No snapshot feature. |
| Server Performance | Varies by cloud provider, some providers have very poor performance; other’s have performance that is the same or better than most dedicated servers. | Performs comparable to most cloud servers. |
| Data redundancy | Varies by provider. Some offer no redundancy some offer same data center redundancy and others have multi- data center redundancy. | Lots of dedicated servers are on a single disk. Some dedicated servers are on a single parity disk (RAID 5) |
| Power redundancy | Cloud servers run on a computing resource pool served by redundant power PDUs, redundant power UPSs, redundant power generators, and redundant power grids (not all data centers were built this way). | Many dedicated servers still sold today have a single power supply. Some dedicated servers have dual power supplies but they are often on a single power PDU. |
| Live hardware resource scaling | Most providers don’t offer on-demand vertical scaling with no-reboot. With some providers you can add CPU cores and RAM without rebooting if the OS supports it. | Upgrading CPU, RAM, and/or disk space requires opening a support ticket shutting down the server, and waiting for the technician to update the hardware configuration resulting in downtime. |
| Hardware upgrade | Some Cloud Computing providers were deisnged in a way that customers are informed that they must move their server and data. Other cloud providers will automatically migrate customers to another server without any downtime. | Dedicated server hosting companies will often allow customers to run on old, out- dated hardware for a long time, when the servers do get replaced the customer is often forced to renew their agreement and select a new server, price and reinstall and reconfigure everything. |
| Costs | $25/month to $199+/month. | $99/month to $1500+month. |
Infrastructure as a Service (IaaS) or cloud hosting is one of the three primary types of Cloud
computing, along with PaaS and SaaS. Many small web hosting companies now offer
Virtual Private Servers (VPS) as Cloud servers and masquerading as “Cloud hosting,” we
thought it would be helpful to clear the air:
A VPS is a traditional single server operating in a multi-tenant hypervisor-controlled
environment. Customers will typically order the VPS with fixed attributes (CPU, RAM,
Storage) in a package and pay a fixed price. Other customers do the same on this dedicated
server, and the provider must make decisions about density – most VPS servers are over-
subscribed. This is why they are so inexpensive and prone to so much variability in
performance. VPS servers are also not fully-networkable – they don’t offer multiple ports
and the ability to connect them to private networks. Most traditional VPS offerings are
billed by the month and have pre-sized configurations.
Pure Cloud computing IaaS offers resources from a group or pool of technology resources
– CPU cores, RAM, Storage and network ports. IaaS users are free to create servers of any
size (if their provider does not limit them to pre-set server or instance sizes). IaaS users are
then able to connect storage devices, firewalls, load balancers and networks (public and
private) together to create a true virtual data center. Cloud computing providers then add
additional features to differentiate themselves, like InfiniBand networking technology that
connects servers to storage and servers to servers at 80Gb/s or “Live Vertical Scaling”
which is where IaaS users can add more RAM or more CPU cores to a server without
restarting a server. IaaS provides far more flexibility, power and isolation that any VPS
platform.
Taking it a step further, Cloud computing IaaS should be thought of as a utility, just like the local electric company, where cloud service providers offer the resources needed to configure virtual servers or instances that do not have fixed computing attributes. The computing configuration of the cloud server can grow or shrink dynamically depending on the resources requested by the specific application or user. Billing for the resources are based on actual consumption over a specific time period for CPU cores, memory, cloud storage, and bandwidth, rather than a preset package that many first generation cloud providers and VPS providers offer. Read more about Cloud Computing in the Cloud Terminology.
Taking it a step further, Cloud computing IaaS should be thought of as a utility, just like the local electric company, where cloud service providers offer the resources needed to configure virtual servers or instances that do not have fixed computing attributes. The computing configuration of the cloud server can grow or shrink dynamically depending on the resources requested by the specific application or user. Billing for the resources are based on actual consumption over a specific time period for CPU cores, memory, cloud storage, and bandwidth, rather than a preset package that many first generation cloud providers and VPS providers offer. Read more about Cloud Computing in the Cloud Terminology.