Picking the right hosting solution for your business can look quite challenging if you’ve just jumped into the whole online world. Since you’re reading this article I can assume that you already had your fair share of experience with various hosting solutions, but as we all know repetition is the mother of learning, therefore I will cover the basics of dedicated server hosting.
So, What Is a Dedicated Server?
Dedicated pronounced as “ /ˈdɛdɪkeɪtɪd/ “, is a word describing something which is “exclusively allocated to or intended for a particular purpose”. Hence, in its simplest form dedicated server can be understood as a type of hosting service, where server/s are allocated just to one client, eg. you.
Unlike shared hosting and VPS hosting services where the resources of a single server (a physical computing machine) are shared among multiple clients, in a dedicated hosting environment all physical resources like CPU, RAM, storage & networking are dedicated just to you.
Dedicated Server Hosting Pros
Naturally, this type of hosting offers many advantages over other hosting solutions.
Even if you have identical resources allocated in virtual or cloud environments, you can get much better performance since all of the server resources are fully allocated to you. There is no middleware between your application and your OS eg. hypervisor, so every component from your dedicated server will be put to work solely for your application needs.
As the only true tenant of that physical machine, you won’t experience any issues caused by other tenants. For example, in virtual and shared environments your service might suffer performance degradation or even downtime due to attacks targeting other clients located on the same physical machine.
Having a dedicated vLAN, isolated CPU & memory environment for your project comes out of the box thanks to being single-tenant. Thanks to this, your security is improved, a lot, but not necessarily per default as a lot of management skills are required to tune up and maintain your server OS security.
You can customize everything. Anything software-related down to the kernel level can be optimized to fit your needs, but what’s even better is that most hosting operators such as ourselves, are giving you the ability to customize your hardware components such as RAM, CPU, storage type, and RAID setup. You can even choose your network interface capacity.
By leveraging the customization part, you can build a very stable machine for your project. For example, you can take a server with a redundant power supply, and build a RAID 10 set, which will give you increased IO performance and make sure that your data is always mirrored, but on top of this, you could have some hot-spare disks that will automatically replace any failed disk drive. In addition, you could allocate multiple NICs which could be connected to different ISPs, giving you the ability to withstand network outages, hardware, and power failures.
Altho the holy grail in the scalable world, these days is horizontal scalability giving you endless opportunities, there is a time when we all prefer simplicity, especially if your project is in its early stage of development. With dedicated server hosting, you can always scale up very easily, without any larger changes in your APP logic. For example, adding RAM, changing your CPU, adding disk, etc. can be done in a couple of minutes, given your server chassis has the physical space to do this. Some can say that you can achieve the same scalability on the cloud even easier, and that will be correct. But nowadays dedicated servers can handle +30TB of RAM, and multiple CPUs with 128 cores, whereas most prominent cloud platforms can offer you 1.4 TB of memory, 96 cores at a much higher cost.
Dedicated Server Hosting Cons
As you can see the advantages of dedicated server hosting can be quite significant, but there are also several downsides with this type of hosting:
It is obvious that all of these benefits come at a cost, but for longer-term projects, we have an option where you can pick some basic configuration that comes at a lower price, and then pay one-time payment fees for various hardware upgrades. Our pricing is crafted in that manner, so your RoI (Return of Investment) will be after four months of service.
This type of hosting requires more skilled persons for its management. But if the cost factor is something that your project can handle, we and many other hosting companies are offering managed servers or server management services, where the management part is taken care of by highly skilled professionals.
What Is an Unmetered Dedicated Server?
As the name “unmetered” (pr. /ˌənˈmēdərd/) suggests, it is something, not measured or assessed by means of a meter. In the case with the dedicated server hosting that would be the network traffic. We have talked about What is Unmetered Hosting, and Why do you need it, I highly recommend you to read that blog post to get a better understanding of the whole unmetered concept.
In short, unmetered dedicated servers are dedicated servers without traffic limits, nor you’re getting charged for some excessive traffic usage. This means that if you get an unmetered dedicated server with 1 Gbps NIC (Network Interface Controller), in theory, you should be able to fully utilize the capacity of that NIC 24/7 without any additional charges.
You may ask yourself how is this even possible, especially when there is an abundance of web hosting providers which are still providing metered services? Well, the thing is that we’re using Tier 4 data centers around the globe to provide our services, these data centers are operated by Tier 1 ISP providers, which gives us a unique advantage to purchase large amounts of bandwidth and use it anywhere in the data centers we use. This gives us the ability to play with our bandwidth consumption on a global scale and given the timezone differences, we’re able to account for the maximum bandwidth usage and hence we’re able to offer unmetered services at very affordable prices. Also by keeping our global bandwidth consumption at around 60-70% we’re able to handle all major spikes in client consumption and yet provide the guaranteed speeds.
Occasionally we have clients using their NIC in full 24/7 but with a large number of clients, these edge situations are insignificant, check out the following graph to gain a better understanding.
Having in mind unmetered doesn’t mean unlimited, if we focus solely on the networking aspect we can see that there is a “limit” after all, and that limit is defined by the bandwidth capacity of the NIC itself, which in this case is 1 Gbps. This means that even if the NIC is fully utilized 24/7 for a whole month you will be able to transfer around 330 TB of traffic.
Here is a table of how much monthly traffic you can generate for a given bandwidth:
|NIC bandwidth||Monthly traffic|
But the NIC bandwidth isn’t the only thing you need to consider, and it does not guarantee that if you have 100 Gbps NIC, you will be able to fully utilize that NIC with any hardware configuration, which leads us to the main point of this blog post.
How to Choose the Best Unmetered Dedicated Server?
When it comes to doing the best choice, its always hard but there are several factors that you should consider prior to making your decision:
Assess Your App Behavior
Figuring out your APP/Project behavior and which resources will be most used is imperative. For example, if your project is a YouTube-like service you will need to account for several factors:
- IO performance
- CPU performance needed for video transcoding
But for comparison, if you’re running live video streaming where no transcoding is involved, then a modern-day average CPU with 16-32 Gb of RAM could serve 1 Gbps without any issues.
Determine Bandwidth Requirements
Determining your project bandwidth consumption can come in handy in the early days of your project development. You can always upgrade your NIC but it may involve some downtime, the majority of our servers are equipped with 1Gbps and 10Gbps NICs so the upgrading process is pretty much seamless, however, upgrading to 100 Gbps might involve short downtime until the NIC card is replaced/attached. So to spare yourself from such situations and avoid paying extra for bandwidth you don’t need at the moment, it is always a good thing to try and estimate your project requirements.
For example, if your YouTube-like website serves full HD video content, that would be 5 Mbps per connection or 25 Mbps for 4k video content, combine this with the simultaneous connections you’re expecting at a given time, and voila you have your estimation.
Choose the Right Hardware
In our opinion, this is one of the most important steps.
Occasionally we do receive requests from clients, complaining that their servers can’t reach the advertised bandwidth, and in 99% of cases is due to not considering the above-mentioned factors, the other 1% is due to bad SFP cable 🙂
The most common mistake is happening due to pre-assumption and not considering other hardware elements, as we’ve mentioned above if you get 10 Gbps that doesn’t guarantee that you will be able to stream videos for your YouTube-like site, stored on a single HDD drive, you see where are we going with this right?
A single HDD drive have running at 5400 rpm can achieve around 90 IOPs, which would result in 3-4MB/s (25-30 Mbps) random read speed, so even if you have 10 Gbps NIC when it comes to reading a video file from the disk it will be fetched with ~30 Mbps from the disk, so you will be able to utilize 30 Mbps of your 10 Gbps NIC. There are other tricks that you can do to improve your app performance, eg. if let’s say you have enough RAM to store the most popular file in it, then in such case, you will be able to utilize your 10 Gbps NIC, otherwise, the HDD will be your bottleneck.
This is the reason why it is very important to analyze your app-specific requirements and figure out any possible hardware bottlenecks.
Through our experiences, we’ve seen various cases of bottlenecks, from storage performance degradation due to the backup process, to various RAM and CPU-related bottlenecks, preventing our clients from fully utilizing their dedicated server’s hardware. That’s why we always encourage our customers to do a quick chat with us, and explain their project so we can share our experience with them and avoid such awkward situations, and of course, we do this for free 🙂
Staying close to your clients it’s very crucial, as this will give you a very good user experience, due to the low latency between your server and your clients, and the overall network connectivity between your service and your clients will be way better as your traffic is routed locally.
Combine these factors and I’m confident that you won’t make a bad decision.
Our Unmetered Dedicated Servers provide great ease of mind when it comes to high bandwidth consumption use cases. Compared to the cloud hosting solutions it will cost you only a fraction of your egress cloud bill. For example, the average price for egress traffic in Europe from top public cloud providers is around $0.05/GB or $50 per 1 TB, compared to our average pricing of $0.001/GB or $1 per 1 TB for our 1 Gbps unmetered servers, definitely, it is worth a thought.
Additionally, you can rest assured that no high bills will be sent to you at any point, due to increased traffic.
I hope you have a better understanding of how to pick your unmetered dedicated server now, and which key moments to pay more attention to in order to get the maximum out of your server.
Alternatively, you can always reach out to me or my coworkers via the support ticket or our live chat and have a chat with us in order to make the right decision.