Table of contents of the article:
Are you tired of hearing about the cloud?
We too a little.
You can blame Google CEO Eric Schmidt for cloud computing. He is credited with having introduced the term in a conference in August 2006.
While network computing has been around since the 60s, Schmidt was the first to launch the cloud as we know it today. You may be collecting a lot of questions from your boss, customers, or development team about why you didn't move to the cloud. After all, who wouldn't want the scalability, redundancy, on-demand services that only the cloud promises?
The cloud offers considerable potential, but few small businesses can leverage what the cloud has to offer. Inflexible operations, inexperience and basic business needs often mean that a dedicated server is the best hosting solution.
If you're not convinced, here are some reasons why you should still use a dedicated server if you don't have enough needs to push you to use a cloud.
Performances
We find that the Dedicated Servers offer the best performance, especially in relation to the price.
We used Rackspace, Softlayer, and AWS. No one can provide the power of a properly configured dedicated server.
This is especially true when it comes to disk I / O - disk reads and writes. With most cloud systems, the network and underlying storage are shared among customers. This can cause disk I / O to be unpredictable. If another customer starts sending large numbers of write requests to the storage array, slowdowns can occur. The upstream network is shared so that bottlenecks can also be experienced.
When troubleshooting performance issues for customers using the cloud or VPS, we generally find disk I / O issues. Often these are not solvable within the cloud framework.
Most cloud providers offer you more storage, not faster storage.
While CPU and RAM can be easily scaled with most cloud vendors, disk I / O cannot often be scaled. While Amazon offers some high disk IO instances, many users continue to build RAID arrays from EBS instances to get the performance they need.
In short, if the operations are relatively simple, a single dedicated server with RAID 10 usually performs faster than more expensive and complex cloud offers, although it should also be noted that normally Cloud storage (at least a quality Cloud like AWS) is significantly more reliable than any RAID solution we wanted to use at the dedicated server level.
Transparency
When debugging performance issues, transparency is key. Transparency is why we are NewRelic fans.
NewRelic allows you to look inside the application and find the bottlenecks. This transparency is the key to solving complicated performance and reliability problems.
Cloud services often obscure hardware and network problems.
- As a shared service, the cloud suffers from two key problems that don't typically occur with i Dedicated Servers.
- Other users have a direct impact on workloads
The underlying hardware errors are often the cause of problems.
With the cloud, you share resources with others. This includes disk, RAM, CPU and network. Cloud software tries to confine and isolate your neighbors, but the fence has holes in it. Often due to inherent design or more often configuration choices, a single user can overload a local compute node. This can cause temporary outages and performance issues for your operations that have nothing to do with you.
Unfortunately, most providers will never recognize or even catch this problem, dragging you to track performance issues.
Hardware errors are another problem. With SoftLayer's service, when we suspect hardware problems on a compute instance, there is no way for us to confirm our suspicions. We simply migrate the instance to another physical node to see if the problem persists.
The cloud makes these migrations easy, but a dedicated server could make these migrations useless. With a dedicated system, we can easily check the hardware and rule out problems. This allows us to focus diagnostic efforts on the right problems.
Redundancy
A common misconception about the cloud is that it is inherently and explicitly redundant. This is often not the case and is not an absolute certainty.
Comparing some solutions to others and calling them all as Cloud is like comparing a Fiat 500 and a Lamborghini Diablo and calling them all cars.
There are cars and automobiles as there are clouds and clouds.
A node in a cloud computing service is usually no more reliable than a single dedicated server.
With cloud computing, the compute node is typically just a server minus the storage. If that node dies, so do your workloads. This isn't much different than a CPU, RAM, or power failure on a dedicated server.
Even with the cloud it is necessary to create redundancy in the system. Switching to AWS won't make your SMB hosting service more reliable or more redundant unless you do it this way.
Check out this setup on AWS. This is very complex and requires significant installation, monitoring and maintenance time.
At the storage level, however, it should once again be reiterated that a serious Cloud solution such as that of AWS (but also of many other vendors) is statistically more reliable than any RAID solution we wanted to install on a dedicated server.
Cost
The cloud costs more. It is obvious, taken for granted and even right that this is the case.
This is true of many small businesses, especially web development and design companies.
Consider a web marketing company that hosts their client's sites. Typically you will have common applications such as WordPress, Joomla, Drupal, and other popular CMSs. You probably also need a hosting control panel like Plesk or cPanel.
When you look at the technical requirements needed to ensure reliable performance for your sites, you will often find that i Dedicated Servers offer you the best quality / price ratio.
The main reason is the lack of redundancy of a dedicated server and therefore a "dry" cost of CPU and RAM.
I often see that VPS or cloud systems are in difficulty with a large number of sites or with high competition precisely because not wanting to spend crazy amounts for the power of instances and resources, we tend to underpower CPU and RAM and therefore create packages of bottle.
It is possible to solve these problems for example by increasing resources significantly or at the disk I / O level for example by creating RAID arrays from storage drives, but this obviously increases costs. And as you add bandwidth, control panels and IP addresses, the cost savings they promised you so much begin to evaporate.
It is difficult to find direct comparisons between cloud and dedicated. Sometimes, you may want to over-deploy the cloud infrastructure to fix some performance issues. Even when you know what you need, the costs haven't been right with many cloud services. Take a look at the AWS “simple” monthly calculator . There is nothing simple there.
Often the costs are based on consumption and not flat, so you cannot have an idea of the traffic nor if a novice developer decides to upload a 20 megabyte PNG rather than a 90 kilobyte JPG and after a week of ADS campaigns on Facebook realize that your customer consumed 50 Terabytes of outbound traffic.
Do you have any idea how much 50 Terabytes of AWS outbound traffic costs?
Calculating a flat rate of 0,080 US dollars per Giga, we have 50 Terabytes, or 50 GBytes multiplied by 0,080, or the beauty of 4000 dollars, about 3800 euros.
Do you know how much our customer spent on 50 Terabytes of traffic that he overstepped using the wrong image format? 30 EUR. Thirty.
Vendor locked locking solutions
Don't get involved.
Being locked out and dependent on a vendor's platform is not. Migration can be painful and costly.
Being stuck and dependent on a supplier's platform can become your ball and chain.
With many cloud providers, if you start integrating more complex services, you may find that you are stuck in their solutions. This can be dangerous if their support, services or pricing changes. Even if the supplier does not change, the technical or business requirements may change. So you need to consider your migration options before selecting a cloud provider.
While the computing side of cloud systems services is generally similar across providers, advanced services such as object-based storage, database abstraction layers, and other technologies often have different APIs. If you build your app to use Amazon's S3, you may need to redesign it to work with another object-based storage model. This can make migration challenging and costly.
Often, I see companies use the cloud provider's advanced services when neither business nor technical needs require such a solution. This creates supplier lock-in where it could be avoided.
I Dedicated Servers they are raw materials. If you are using a hosting control panel such as Plesk or cPanel, migrating to another server or service provider is a simple and well-documented process.
So when a dedicated server can meet your business and technical needs, why risk vendor lock-in?
Scalability
You are not ready to resize.
One of the main marketing points for cloud services is scalability. Although compute resources can be resized, applications or operations may not be ready to be resized.
If you are using a hosting control panel, the scalability options are limited. You can increase your CPU / RAM or add a dedicated database, but you already have these options with Dedicated Servers. The cloud makes it easier.
As mentioned above, scaling disk I / O is often not available or limited with the cloud. In our performance tuning work, disk I / O is often the main performance issue, particularly with shared hosting operations.
Don't be fooled by advertising. You can't just plant your operations in a cloud provider's backyard and expect it to magically grow.
Applications need to be built and managed with scalability in mind. Attempting to cram legacy applications into a modern, scalable cloud framework often results in failure.
Also, ask yourself why you need to resize anyway?
If your websites are slow, perhaps optimizing your server configuration and eliminating application bottlenecks will solve the problem. The cloud will not solve fundamental programming inefficiencies.
Go Peace of Mind with a Dedicated Server
If you are a small business with relatively simple hosting operations, then don't ignore the Dedicated Servers. I know the pressure from customers and vendors to use the cloud is powerful, but that's because they only see the marketing exaggeration.
The reality is that a properly managed dedicated server will generally provide higher performance and reliability at a lower cost than current cloud service options.
Ask for advice
We are used to evaluating the best solutions for our customers every day. Sometimes the Cloud solution is the only option, many others i Dedicated Servers (perhaps redundant in clusters) are the only solutions if you expect performance and be able to stay within a budget, however reduced.
Propose your business needs to us and you will see that we will be able to explain in a very eloquent way why you should use one solution rather than the other. In short, do not get carried away by fashions, but let yourself be advised by those who, like us, now have numerous examples of cases under their belt.
