ThankReview evaluated and assessed 22 VPS hosting plans ranging from $60 to $100 in the previous 12 months. We assessed their CPU power, network performance, disk speed, web performance, and stability.

How the scores are calculated

Each of the five categories is graded: web, CPU, disk, network, and stability. Each category grade is composed of one or more metrics selected from appropriate test kinds. Given a plan and a test measure, the grade for that metric is calculated relative to the same metric from all other plans. A plan with a measure in the top 18% of all plans receives an A, while a plan with a metric in the bottom 18% receives an F. A grade is a numeric value between 0.0 and 6.0 under the hood. The total score is the average of all grades placed into the [0 – 100] range. The price of the VPS package is not taken into account while calculating the grades.

All VPS scores we tested in one chart

In the figure below, each horizontal bar segment reflects a grade value: A: [16-20], F: [0-3]. The Score is the overall length of the bars.

ThankReview evaluated and assessed 22 VPS hosting plans ranging from $40 to $60 in the previous 12 months. We assessed their CPU power, network performance, disk speed, web performance, and stability.

How the scores are calculated

Each of the five categories is graded: web, CPU, disk, network, and stability. Each category grade is composed of one or more metrics selected from appropriate test kinds. Given a plan and a test measure, the grade for that metric is calculated relative to the same metric from all other plans. A plan with a measure in the top 18% of all plans receives an A, while a plan with a metric in the bottom 18% receives an F. A grade is a numeric value between 0.0 and 6.0 under the hood. The total score is the average of all grades placed into the [0 – 100] range. The price of the VPS package is not taken into account while calculating the grades.

All VPS scores we tested in one chart

In the figure below, each horizontal bar segment reflects a grade value: A: [16-20], F: [0-3]. The Score is the overall length of the bars.

ThankReview evaluated and assessed 31 VPS hosting plans ranging from $25 to $40 in the previous 12 months. We assessed their CPU power, network performance, disk speed, web performance, and stability.

How the scores are calculated

Each of the five categories is graded: web, CPU, disk, network, and stability. Each category grade is composed of one or more metrics selected from appropriate test kinds. Given a plan and a test measure, the grade for that metric is calculated relative to the same metric from all other plans. A plan with a measure in the top 18% of all plans receives an A, while a plan with a metric in the bottom 18% receives an F. A grade is a numeric value between 0.0 and 6.0 under the hood. The total score is the average of all grades placed into the [0 – 100] range. The price of the VPS package is not taken into account while calculating the grades.

All VPS scores we tested in one chart

In the figure below, each horizontal bar segment reflects a grade value: A: [16-20], F: [0-3]. The Score is the overall length of the bars.

ThankReview evaluated and assessed 43 VPS hosting plans ranging from $10 to $25 in the previous 12 months. We assessed their CPU power, network performance, disk speed, web performance, and stability.

How the scores are calculated

Each of the five categories is graded: web, CPU, disk, network, and stability. Each category grade is composed of one or more metrics selected from appropriate test kinds. Given a plan and a test measure, the grade for that metric is calculated relative to the same metric from all other plans. A plan with a measure in the top 18% of all plans receives an A, while a plan with a metric in the bottom 18% receives an F. A grade is a numeric value between 0.0 and 6.0 under the hood. The total score is the average of all grades placed into the [0 – 100] range. The price of the VPS package is not taken into account while calculating the grades.

All VPS scores we tested in one chart

In the figure below, each horizontal bar segment reflects a grade value: A: [16-20], F: [0-3]. The Score is the overall length of the bars.

ThankReview evaluated and assessed 28 VPS hosting plans ranging from $5 to $10 in the previous 12 months. We assessed their CPU power, network performance, disk speed, web performance, and stability.

How the scores are calculated

Each of the five categories is graded: web, CPU, disk, network, and stability. Each category grade is composed of one or more metrics selected from appropriate test kinds. Given a plan and a test measure, the grade for that metric is calculated relative to the same metric from all other plans. A plan with a measure in the top 18% of all plans receives an A, while a plan with a metric in the bottom 18% receives an F. A grade is a numeric value between 0.0 and 6.0 under the hood. The total score is the average of all grades placed into the [0 – 100] range. The price of the VPS package is not taken into account while calculating the grades.

All VPS scores we tested in one chart

In the figure below, each horizontal bar segment reflects a grade value: A: [16-20], F: [0-3]. The Score is the overall length of the bars.

In latest 12 months, ThankReview placed 180 web servers and cloud servers from 40 providers through a rigorous series of benchmarks to assess their performance, capability to handle large loads, reliability, and network speed. These are the best virtual private servers in six pricing levels.

Introduction about the winners

Vultr, situated in the United States, has some of the greatest management UI, a wide range of functionality, and global presence.

Genesis Hosting Solutions is a US-based provider with a datacenter in Illinois, a diverse product offering, a comprehensive collection of features, and some of the most cutting-edge technology available in the VPS market.

Contabo is a European datacenter provider with datacenters in Germany, the United States, and Singapore. It provides both high-end cloud servers (VDS) and low-cost servers with high specifications (VPS). It has lately begun to provide Object Storage.

UpCloud is a rapidly growing European provider with 12 datacenters throughout the world, a robust feature set, and strong and consistent performance.

IDrive is an American cloud provider with a strong core business in cloud backup solutions that will grow into compute solutions in 2021. It provides a robust product range of cloud servers in 20 locations.

OnetSolutions and Alwyzon are tiny providers situated in France, Estonia, and Austria, respectively, with a simple set of features but excellent performance and affordable pricing.

Layerstack is a Singapore-based service with data centers in four locations across Southeast Asia and America. Layerstack is the first provider at VPSB to use EPYC Milan CPUs.

OVHcloud is a well-established major provider with a robust feature set, many datacenters, and a diverse product portfolio that allows everyone to find what they want.

Best virtual private servers in six pricing levels

ThankReview evaluated and assessed 16 VPS hosting plans ranging from $0 to $5 in the previous 12 months. We assessed their CPU power, network performance, disk speed, web performance, and stability.

How the scores are calculated

Each of the five categories is graded: web, CPU, disk, network, and stability. Each category grade is composed of one or more metrics selected from appropriate test kinds. Given a plan and a test measure, the grade for that metric is calculated relative to the same metric from all other plans. A plan with a measure in the top 18% of all plans receives an A, while a plan with a metric in the bottom 18% receives an F. A grade is a numeric value between 0.0 and 6.0 under the hood. The total score is the average of all grades placed into the [0 – 100] range. The price of the VPS package is not taken into account while calculating the grades.

All VPS scores we tested in one chart

In the figure below, each horizontal bar segment reflects a grade value: A: [16-20], F: [0-3]. The Score is the overall length of the bars.

All marks are based on ThankReview tests that are done on each VPS plan. Web, sysbench, endurance, remote timings, and network transfers are the five types of testing. Each grade is made up of one or more test measures.

Each of the five categories is graded: web, CPU, disk, network, and stability. Each category grade is composed of one or more metrics selected from appropriate test kinds.

Given a plan and a test measure, the grade for that metric is calculated relative to the same metric from all other plans. A plan with a measure in the top 18% of all plans receives an A, while a plan with a metric in the bottom 18% receives an F. Relevant single metric ratings are then weighted and added together to form a category grade.

A grade is a numeric value between 0.0 and 6.0 under the hood.

The total score is the average of all grades placed into the [0 – 100] range.

The price of the VPS package is not taken into account while calculating the grades. More costly plans with more specifications are more likely to receive higher marks than less expensive ones. Turning on the “price weighted” switch in the Screener will modify this.

Web Performance
The Web Performance grade is based on the average response time of the Web test, the 99th percentile response time, and the highest rate of requests without error.
Raw CPU Power
The Raw CPU Capacity grade is calculated using the Sysbench CPU test number of operations per second and the Endurance test number of iterations per hour.
Performance Stability
The Web test average reaction time standard deviation, the difference between the 99th percentile and average response times, and the Endurance test number of iterations per hour standard deviation all contribute to the Performance Stability grade.
Disk IO Performance
The Sysbench test sequential disk write speed and random read/write performance are used to calculate the Disk IO Performance grade.
Network Performance
Large file network upload and download transfer speeds comprise the Network Performance grade.

Grades Based on Price

The option to display “price weighted” grades is available on the Screener and other pages. With this option enabled, a low-cost plan with the same performance as a high-cost plan will be displayed with better grades.

The price weighted perspective attempts to address the following questions:

Which plans offer the best value?

What kind of return do I receive for every dollar I spend?

For low-cost plans, the “bonus” grade is equivalent to the difference between the $30 and monthly prices. The bonus calculation algorithm is intended to have a minimal influence on extremely low grades and a maximum impact on mid to high grades.

Similarly, when this option is enabled, grades for high-priced plans are displayed lower: the more the monthly price is above $30, the greater the knock on the grades. In this scenario, the function is intended to have a minimal influence on great scores while having a significant impact on medium to low grades.

We are frequently asked which server provider we suggest for hosting WordPress in combination with ThankReview. Because ThankReview works with any and all providers, it might be difficult to pick from the hundreds, if not thousands, of available suppliers. I’ll go through five of the most popular in this article: DigitalOcean, Linode, Vultr, Amazon Web Services (AWS), and Google Cloud.

Although both AWS and Google Cloud provide a full range of cloud computing services, we’ll focus on which provider is ideal for running WordPress on a single server, without taking into account other services like load balancers, managed databases, or object storage. We know that because we’ll be utilizing ThankReview to provide the servers, all five will be set to the identical specifications. As a result, while evaluating the five services, we can focus on simplicity of use, performance, and pricing.

In this essay, I’ll also be utilizing Amazon Lightsail. Amazon Lightsail removes many of AWS’s complexity and makes it simple to set up virtual servers while still utilizing the same underlying AWS infrastructure. While DigitalOcean, Linode, Vultr, and Amazon Lightsail have equivalent server specs, Google Cloud does not, thus I’ve chosen the closest spec VM instance, which is a g1-small at the time of writing. I’ve also added a comparable size SSD boot drive because all other suppliers include one as standard, giving us the following servers:

Performance

I’m going to provision a server on each of the following server providers to test their performance. Because we’re utilizing ThankReview, all five servers will be optimized for WordPress and run the same server stack:

• Ubuntu 20.04
• Nginx 1.18
• PHP 7.4
• MySQL 8.0

It’s time for some benchmarking! On each server, I’ll set up a single site with a fresh install of WordPress running over HTTPS:

• WordPress 5.5.3
• Twenty Twenty 1.5

Cloudflare hosts the DNS for all five sites, with proxying Cloudflare hosts the DNS for all five sites, with proxying disabled. The default PHP OPcache settings are in effect, and a persistent object cache is set up. Each test will be repeated five times, with the average of the five results used for comparison. I’m running the benchmarks with Loader, which sends queries from North Virginia. All of our test servers are situated in data centers in the US east (save for Northern Virginia), which should produce unbiased findings.

Cached Performance

When full page caching is enabled, this test mimics a spike of traffic hitting the site’s homepage. We’ll send 50,000 customers to each site over the course of 30 seconds and record the average response time (lower is better). The cache has already been pre-warmed (meaning all requests will hit the cached version).

Uncached Performance

This test simulates a higher continuous demand with page cache disabled, which puts a greater strain on the server because Nginx, PHP, and MySQL are all competing for server resources. We’ll send 10,000 customers to each site over the course of five minutes and record the average response time (again, lower is better).

Response times are substantially faster this time, which is to be anticipated given that each request must query the database and display the page in PHP before being delivered to the client. Linode’s response times were significantly slower than those of the other providers, so I ran the benchmarks again the next day. However, the outcomes were comparable.

Ease of Use

Let’s start with how simple it is to set up each service on ThankReview.

DigitalOcean

The target audience for DigitalOcean has always been developers: designed for developers: Designed for developers. Build more and spend less time managing your infrastructure with our easy-to-use control panel and API.

They took the relatively complex process of deploying servers to the cloud and simplified it. Their user interface is stunning, and cloud servers can be set up in minutes. Pricing has always been transparent, and you know exactly how much you’ll be paying each month.

In many ways, DigitalOcean is similar to Stripe for online payments, which is well-known for its user experience, documentation, and API.

For these reasons, it was an obvious choice for DigitalOcean to be the first server provider integrated directly into ThankReview. This makes it simple to spin up DigitalOcean servers through ThankReview. You simply select the server size and region after entering your DigitalOcean API credentials, and ThankReview handles the rest without requiring you to leave the ThankReview control panel.

Amazon Web Services (AWS) and Amazon Lightsail

AWS offers a full range of cloud services (over 165 at the time of writing) that are all handled through their AWS dashboard. Unlike the other providers, AWS is not built for ease of use, and the AWS console may be daunting. Pricing is even more ambiguous because services are charged by the hour or consumption (depending on the service). The fact that AWS offers a certification program as well as a pricing calculator should give you an indication of the complications involved.

Virtual servers are known as Elastic Compute Cloud (EC2) instances in the AWS ecosystem and require quite a bit of additional configuration to get up and operating.

Enter Amazon Lightsail, a user-friendly UI wrapper for several of their most regularly used services. Similar to DigitalOcean, Amazon Lightsail makes it extremely simple to build cloud servers. Pricing is also simplified, with each server’s monthly cost shown.

Although it is not as simple as adding an AWS EC2 or Amazon Lightsail instance to ThankReview, it is still doable. Simply provide your instance’s IP address, username, and authentication method of choice. Do you require further assistance? Check out our Amazon Lightsail documentation.

Google Cloud

Google Cloud is comparable to AWS in that it provides a comprehensive set of cloud services. Although Google Cloud is not as simple as DigitalOcean or Amazon Lightsail, it is easier to use than installing AWS EC2 instances. However, Google Cloud lags behind in terms of cost. Google Cloud’s price is not only significantly more than that of the other services, but it is also less transparent. When launching a VM instance, you must depend on the pricing estimate.

Google Cloud VM instances, like Amazon Lightsail, may be added to ThankReview by giving the instance with the necessary connection data. Do you require further assistance? Check out our Google Cloud documentation.

Linode

Linode, like DigitalOcean, is geared for developers who want to simplify their cloud infrastructure: Simplify your cloud infrastructure with our Linux virtual machines and robust set of tools to develop, deploy, and scale your modern applications faster and easier.

Linode entered the market a few years before DigitalOcean and quickly earned a large following, owing to its minimal competition and good tools and instructions. However, their control panel received little attention over the years, and it gradually slipped behind DigitalOcean in terms of appearance and usability. They’ve gradually brought out their newer, more contemporary cloud management over the last few years, but it still lags behind DigitalOcean.

Having said that, Linode is still a fantastic company, with fairly priced servers and straightforward pricing. When accessing their cloud management, don’t anticipate the same degree of refinement as DigitalOcean.

ThankReview can also use Linode servers. Do you require further assistance? See our How to Install WordPress on Linode doc for more information.

Vultr

Vultr is a relatively new VPS service in comparison to the competition, but they’ve grown in popularity in recent years. Their control panel is comparable to DigitalOcean, and their high frequency compute instances are outstanding value for money (more on that below).

Vultr servers, like Linode, may be utilized with ThankReview. See our How to Install WordPress on Vultr guide for more information.

CPU-Optimized Servers

We recommend utilizing a CPU-optimized server for hosting sites where page caching is not possible (think dynamic WordPress sites like BuddyPress or WooCommerce). Because of the amount of moving elements involved in accepting an incoming request, handling an uncached request is substantially more demanding on server resources, as we’ve seen above. CPU-optimized servers employ high-performance hardware, which allows them to respond to requests more rapidly.

Except for Amazon Lightsail, all of the following providers provide a CPU-optimized server. However, AWS EC2 also provide CPU-optimized servers, so we’ll compare those instead of Amazon Lightsail. For comparison, I’ve chosen the following servers:

We’ll run the same test that we performed for uncached performance, which involves sending 10,000 customers to each site over the course of 5 minutes and measuring the average response time (again, lower is better).

Conclusion

I’ve always been a big fan of DigitalOcean (and I still am). I like the approach they’ve adopted over the years to make it easier for developers like me to create cloud servers. Not only that, but I’ve always found their pricing to be both reasonable and simple. I hope they continue to improve the cloud experience for non-DevOps professionals. Look no farther than DigitalOcean for the most easy server provider to utilize with ThankReview. They’re growing into other cloud providers, like AWS and Google Cloud, and have just released managed databases and virtual private clouds. I’m looking forward to seeing how they simplify the cloud experience in the future.

Until the release of Amazon Lightsail, I had avoided utilizing AWS for anything other than S3 and Cloudfront. The sheer amount of AWS services makes choose which service to utilize for a certain use case challenging. Because of the confusing UI and cost, AWS has never been my first choice for cloud servers. However, Amazon Lightsail is a significant step forward. Although the UI and UX aren’t nearly as good as those of DigitalOcean or Vultr, it’s far superior to creating EC2 instances through the AWS interface.

Despite the fact that I’ve been using DigitalOcean for a few years, perhaps it’s time to give Vultr a go the next time I need a cloud server. Their user interface and user experience are comparable to DigitalOcean, and their high frequency compute instances provide unrivaled performance for the pricing.