Cost is an important factor to consider when evaluating different VPC providers, but the kind of hardware and data lines the providers offer plays a role, too.
There are many different virtual private cloud options available, but how can you choose between one provider and the next? Should you go with the most well-known, or the provider with the best price? Or should you base your decision on something else entirely?
Selecting a virtual private cloud server provider might not be as simple as it seems. Every provider's offerings are different, so there are a number of different criteria you should take into consideration before making a decision.
What is the cost?
In almost every IT environment, there is constant pressure to keep costs in check, so service cost might be one of the first things you want to consider. But estimating the cost of a virtual private cloud is notoriously difficult.
Providers typically don't charge subscribers for the virtual private cloud itself; they bill subscribers based on the resources they use in the public cloud. But these resources aren't usually billed at a flat rate. For example, there might be a predictable hourly charge for running a particular VM instance in a virtual private cloud, but there are almost always ancillary charges for bandwidth and storage consumption, as well as the use of other back-end resources.
Amazon provides a cost calculator to help AWS subscribers estimate the cost of running workloads in the AWS cloud, and some other providers offer similar tools. These calculators might be your best option for performing head-to-head cost comparisons of virtual private cloud server providers.
Can you get dedicated hardware?
Another consideration to take into account is whether your virtual private cloud will be based on dedicated hardware or if your organization will share hardware with other tenants.
Hardware sharing helps keep prices low because your organization splits the costs with other organizations. But the problem with shared hardware is noisy neighbor syndrome; if another tenant is allowed to consume excessive resources, it could negatively affect your workload's performance.
Each virtual private cloud server provider has its own way of doing things. Virtual private clouds often use shared hardware by default, but most providers give their subscribers the option of using dedicated hardware for an additional cost.
What is the underlying platform?
Another consideration is the underlying platform. For example, an organization that is heavily invested in on-premises VMware infrastructure should probably consider a virtual private cloud server provider that also uses VMware. Doing so might create a consistent management experience both on premises and in the public cloud depending on how much of the infrastructure the provider actually exposes to tenants.
Does the provider offer dedicated data lines?
Another thing to consider is whether or not the private cloud server provider offers dedicated connectivity or if you must connect to the service using a virtual private network.
A dedicated connection to a virtual private cloud server provider tends to perform better than basic internet connectivity because there is less overhead because virtual private cloud traffic doesn't have to compete with internet traffic for bandwidth.
In smaller organizations, dedicated connectivity probably won't matter much. Using dedicated connectivity tends to be far more expensive than connecting over the internet, and the cost is likely to be beyond the reach of SMBs.
In the case of larger organizations, however, dedicated connectivity is definitely something to consider. A dedicated connection to a virtual private cloud server provider tends to perform better than basic internet connectivity because there is less overhead because virtual private cloud traffic doesn't have to compete with internet traffic for bandwidth. As a bonus, offloading virtual private cloud traffic to a dedicated connection can improve internet performance, as well.
Are you limited to a single subnet?
While you are objectively comparing virtual cloud server provider offerings, you should also take into account whether the environment can accommodate multiple subnets or if the provider limits you to using a single subnet.
The major vendors generally enable you to create multiple subnets if you have the need, but the same is not necessarily true of all the boutique providers. As such, it's a good idea to check subnet limitations.
Is there support for multiple VPCs with separate security domains?
Finally, it's very important to check whether the provider will allow you to create multiple virtual private clouds.
When an organization is first getting started with an IaaS cloud, it might at first create a virtual private cloud and use it to offload some of its on-premises resources, or use it to create some new workloads. But, over time, organizations commonly realize they need to establish isolation boundaries between certain workloads, or perhaps even give certain business units their own dedicated virtual private cloud.
This brings up another important point. It isn't enough for a virtual private cloud server provider to enable subscribers to create additional virtual private clouds. The infrastructure must also support the use of separate security domains. Otherwise, providing users with a self-service experience similar to what they would get in a private cloud won't be an option.
