When do you replace or upgrade your server? These are our often seen top 5 server replacement considerations.
- Your server is getting old and the hardware is showing it’s age
- Your operating system software will no longer be supported or is obsolete
- You need to up size (or downsize) because your requirements have changed
- You’re expanding and opening up a branch location
- Your equipment lease is coming due
Your server is getting old and the hardware is showing it’s age
This is the most common scenario we hear from customers. When your server gets old it becomes much more likely that a hardware failure will occur that can take the server offline. Servers and their components fail, there’s no sugar coating it. Its a fact that you will have to replace your server due to hardware failure, and it’s not a question if if, but rather when.
As with all things, use them long enough and you’ll wear them out. Consider that servers run 24x7x365, rain, shine or otherwise. After 5 years of service a server has 43,800 hours on it! That may or may not seem like a lot to most people, but it’s many, many times the number of hours you accumulate on your car or your furnace, or your air conditioner. Even your fridge doesn’t have this amount of wear on it! Our appliances don’t accumulate the wear on them because they cycle on/off. Servers don’t have that luxury.
Sure, servers have components that are robust and will last a very long time. But there are also mechanical parts in them, and they are subject to an often dusty and dirty environment. These parts accumulate dirt and debris and can lead to persistent overheating conditions and premature failure.
Lets be real here, chances are you have never powered down your server and cleaned it to help address overheating issues! Replacing your server due to age is one of the things you can do to avoid downtime when you least expect it.
Often replacing the hardware will allow you to do an operating system change as there are usually discounts available when purchasing the two as a bundle.
Your operating system software will no longer be supported or is obsolete
This one makes it to the top 5 of our common server replacement considerations. Usually obsolescence happens once every 8-10 years and mainly this is because most customers run Microsoft Windows operating systems. Microsoft is fairly clear that their OS’ will be supported for up to 10 years, and that means your server gets 10 years of hardware drivers, updates, security patches etc.
In addition to this, 10 years for an operating system is an eternity! To take advantage of the new technology built into new versions of software like SharePoint, Exchange, SQL Server, etc. you’ll need to stay current with your operating system.
Microsoft has done an amazing job of maintaining backwards compatibility, but even they cannot put a 2021 Mustang engine into a 1970’s Ford Pinto! Many of the issues Microsoft runs into are due to the changing landscape around how we use technology and in some cases previous generation operating system architecture is just not designed to be as flexible as we need it to be.
Pop over to this site to check Microsoft’s end of support schedule.
No sugar coating this one either they’re also in the business to make money. If their 20 year old operating systems are still relevant and supported why would people change?
Often it makes sense to change the server hardware when you’re forced to do an operating system change. This does depend on how old the hardware is, but you can usually get some great discounts on the OS when you bundle it in with the equipment.
You need to up size (or downsize) because your requirements have changed
This one ties into the next top 5 server replacement considerations. As time passes your situation changes and the server you have is not up to the job. Or another more rare, but not uncommon situation is that you’ve sold off a part of your company and the server is leaving as part of the deal.
In the case of up sizing, I would recommend keeping the existing server if it’s not end of life. Many times having a spare allows you to implement fail over clustering, load balancing and high availability. Sure the old server is going to be much slower, but it may allow you the opportunity to survive what could have been an extended period of downtime without much additional investment.
The other thing to consider is that any secondary server in a fail over/high availability environment generally does not see the load and subsequently the wear that the primary server does. Of course this goes out the window when we talk about load balancing, but that scenario also allows you to improve your applications responsiveness overall as two servers will be faster than just one.
In addition as you scale out your server infrastructure, any single “node” (server) failure has a decreasing impact to your overall infrastructure performance. For example, if you only have one server, you’re at capacity if that server is at capacity and you’re also down if that server fails. Two servers, and you’re at capacity if both are, but if any one fails you’re still operating, but with a 50% reduction to your capacity. Three servers and any one server failure reduces your capacity to 66% and so on.
I would advocate against “replacement” of servers when the one being discussed is still functional and the OS is still supported. You need to remove equipment due to age and software obsolescence but first consider adding another server to your environment instead of “replacement.” A good analogy – you need a new car, but the old one still works. Consider keeping it (if it doesn’t add substantially to your costs) to cover off those times where the new car is unavailable or down for repairs.
If you have 4+ redundant servers you’re probably over capacity so any additional servers will not add much resiliency, so replace the oldest as you go. If you only have one server and are thinking of replacing it, I’d firstly make an evaluation to determine if it can be used for redundancy. It gives you so much more flexibility and resiliency. A big advantage of adding just one redundant server is that you gain the ability to do many after hours maintenance tasks during the workday. An example: patching and updating your servers can happen more often during the day and special arrangements don’t need to be made for after hours work.
You’re expanding and opening up a branch location
Before you consider adding a server to your branch office location, I would consider looking at:
- Is there capacity in your existing data center for the additional office?
- Do you already have two redundant sites for your infrastructure?
- Does a switch to a cloud provider make more sense?
When a company has grown to the point that it is considering it’s first branch office consider it an opportunity. It’s an opportunity to review existing infrastructure and leverage that new location for geographical redundancy.
As you grow, you’ll want to expand and reduce your risk of downtime. Being dependent on one single location is one of these risks. Let’s imagine a scenario where you have all your infrastructure in your head office in Edmonton and you have a satellite office in Calgary. If a power outage occurs in Edmonton, your company is down and everyone cannot function. If a power outage happens in Calgary, Edmonton can still operate. You’ll want both Edmonton and Calgary to operate regardless of which site is down.
Before putting a new server in your branch location, examine your infrastructure. If you already have two locations consider improvements on your branch office internet services instead of new server gear. If you have infrastructure in your head office and this is your first branch location, perhaps relocating some servers from the head office is a better choice. That will give you that redundancy, resiliency, and performance instead of adding additional equipment.
I know IT folks like to spend money and 95% of the time it can be justified. We believe this as a situation where they should get everything they need, but not necessarily everything they want. Branch offices are usually quite small (not always!). Consequently your cost per user supported shouldn’t disproportionately higher than at the head office. Whatever equipment you put in that branch office should have value for the entire company and not just that branch office.
Your first branch location can be leveraged as a fantastic way to reduce IT risk across your company. Evaluate your options to see what fits best.
Your equipment lease is coming due
What do you do? you have 2 months left on your lease and you have a decision to make. Buy it out? replace (renew with new terms)? go to the cloud?
This is a complex issue and entirely depends on:
- Cost to transition to new gear, or the cloud
- CAPX vs OPEX
- What are the buyout terms?
- How many years of life does the equipment have left?
- What does technology look like in 5 years?
A simple blog posting isn’t going to get you the answers you want. Our advise is do your homework by talking to your IT and business development teams. Of course speaking with your accounting team goes without saying!
I would consider changing equipment if the lease was 5 years or longer. In fact I wouldn’t lease gear for a longer than a 5 year term! You risk the equipment becoming obsolete, and potentially discarded long before it’s paid for. Look at the gear like you would a car (depreciating asset with finite lifespan).
If the buyout is minimal, it usually makes sense to exercise the option as discussed above regarding redundancy and resiliency. New equipment can be leased on an as needed basis going forward.
A well known secret – it’s usually in your best interests to exercise the buyout option on computer equipment. Buyouts are often very favorable as technology equipment has very little value at the end of it’s life. What would you pay for a 5 year old obsolete server with 43,000+ hours on it? Leasing companies will auction off the old equipment and make some money but overall it’s not much at all. This is unlike a 5 year lease on a car which will have a residual value when the lease expires.
I hope you’ve found this article informative! And if you are considering replacing your server give us a call we can help clear up any unknowns and walk you through what a cloud option might look like.
Need more information? no problem. contact us today
Please follow our FB and LI pages for more information like this and be sure to check back from time to time.