Selecting your server: A guide to the most critical components


August 18, 2016

Your business is growing every day, and that QuickBooks database just isn’t pulling its weight anymore. Opening files used to be instant but now feels like waiting on the highway in rush hour. Looking at that 6-year-old desktop computer covered in dust in the corner, you think, maybe it’s time to upgrade. Once you do some research, you decide that you want to get a server. Congrats on making the choice to take your company to the next level! However, when you begin to encounter some of the intimidating technical acronyms out there, you realize how deep the technology rabbit hole goes and how shallow your pool of knowledge really is.

Although servers offer a lot of useful technology that is critical for maintaining business continuity and generating a return on investment, the learning process can be very time-consuming. But don’t worry! As an IT expert, I am here to shed some light and provide some recommendations to help ensure you make the right server choice.

Before you begin

The first thing you need to do is to create a list of software you will be running on the server, e.g. Quickbooks, SQL Server, Microsoft Exchange, and Terminal Server. You should also note their specifications, such as the RAM, CPU, and HDD storage space recommended to run. When selecting your server, these are the basic requirements that you will want to exceed.

Selecting a model

Model (also known as chassis) selection is largely dependent on whether you have a server rack. If your server room has a two- or four-post rack in it, I recommend:

  • Dell PowerEdge R330: This one-CPU socket server is great for serving files and hosting applications.
  • Dell PowerEdge R430: This two-CPU socket server can hold up to more intensive applications, like SQL, Exchange, and Remote Desktop/Terminal Server.

If you don’t have a server rack, go with the Dell PowerEdge T430.

Hard drive configuration

When done incorrectly, the nuanced art of hard drive configuration can negatively impact performance and future upgrade capabilities. Below, I outline each key feature, along with my recommendations.

  • 5 or 3.5 form factor: Hard drive form factor should almost always be 3.5 unless you are building a very storage-dense server.
  • Hot-Plug/Hot-Swap or cabled: Hot-plug drives are a godsend when hard drive failure occurs. This feature allows you to remove and replace a hard drive while the server remains online and functioning, provided you have the appropriate RAID configured (more on that later). Always get this, period.
    • SATA is a low-end hard drive technology. It is slower and has a lower duty cycle (meaning the reading/writing expectations for the drive are low), but is also the least expensive and offers the highest storage capacity per drive. SATA should be reserved for low read/write file servers, as the drives will wear out fast if they are constantly in use.
    • NLSAS is a mid-range hard drive technology. It has the rotational speed of SATA, but is rated for a high-duty cycle (usually 100%) and communicates using the SAS protocol. As with SATA, you can get NLSAS drives in high-drive capacities, but NLSAS is much more reliable. I advise storing files and other large data on NLSAS drives.
    • SAS is the high-end hard drive technology. It offers superior rotational speed and 100% duty cycle ratings, but the drives tend to be much lower in capacity than SATA or NLSAS. SAS drives should be used for high read/write applications and operating system boot drives.
    • SSDs have just started to gain traction in the enterprise world. They offer unparalleled speed for data read/write, but at a very steep cost. It is best to avoid SSDs until the technology becomes cheaper to implement.

So, how many drives should you buy? I typically leave one or two bays open when making a new server to ensure expandability in the future. Following this logic, simply take the number of hard drive bays on the chassis and subtract one or two.


RAID is a technology for combining multiple physical hard drives together as one storage pool. Certain RAID levels even allow you to replace failed hard drives without losing data. This is tied in closely with the hard drive configuration but is unique enough to get its own section.

You’re dealing with two major components here: RAID controllers (which control the writing of data to each drive) and RAID level (the logic behind how you want the RAID to operate).

  • RAID controllers: Hardware RAID controllers like the Dell PERC H730 are my go-to’s. They offer vastly superior queue depth and caching than the budget-model H330. The cost difference is only around $200, and the performance gain is immense.
  • RAID level: There are many different RAID levels (0, 1, 5, 6, and 10 are the most common) and each has its own advantages and disadvantages. The simplest setup you can use to minimize costs and provide failure tolerance is a RAID 5, which allows multiple drives to be combined as a storage pool while tolerating one drive failure at a time without losing data (just make sure to replace the drive as soon as it fails).

NOTE: Putting hard drives in a RAID array will cause your usable storage space to be less than the total drive capacities. Use RAID calculators like the one HERE to see what your storage pool looks like after the RAID.


The CPU acts as the brain of your server, so you’ll want to spend a bit more money on this component if possible. Keep in mind that it is much more costly to upgrade your CPU later (both monetarily and in terms of server downtime) than it is to buy a high-grade CPU from the start. I advise purchasing the 6-core or 8-core versions of the Intel Xeon E5, as they offer very high performance in a single CPU at a moderate price. However, you should refer to your application’s recommended specifications to ensure this is appropriate.


RAM is what enables your server to multi-task the jobs you throw at it. I advise using a minimum of 16GB of RAM, although most motherboards offer enough RAM slots to make upgrading later a non-issue. Again, you should refer to your application’s recommended specifications to ensure this is appropriate.

The bottom line

Buying a server is a big corporate investment, and the elements outlined above are the main components in terms of making sure your server stays online and operating. Because buying a server is such a mission-critical decision, we encourage you to reach out to an established managed IT services provider to help guide you through the selection process.

We're Integris. We're always working to empower people through technology.

Keep reading

4 Cybersecurity Takeaways from China’s Largest Data Breach

4 Cybersecurity Takeaways from China’s Largest Data Breach

Cybersecurity drama strikes again as human error leads to China's biggest data breach and perhaps the most significant hack of personal information in history. According to Threat Post, the incident was triggered after a Chinese government software developer wrote a...