The good, bad and ugly of on-prem.
Is this the right choice for your organisation?
Colo? On-prem? Cloud? Hybrid? There are so many options and questions that follow. To help cut the complexity out of the decision making process, we took a warts-and-all look at all of the options available to you, when it comes to where you store your data.
Your data centre is one of your company’s most strategic assets. It’s the glue for your critical infrastructure that connects your business to your data, applications, staff, customers, and suppliers.
Finding the right IT solutions for your company is complex enough, but the pandemic we’re all living through and the rapidly shifting sands surrounding fluctuating market conditions and customer behaviour complicate things further.
The trick is to find a solution that can support your needs now, and will scale easily with you as your requirements evolve. Many IT teams are actively shifting workloads to cloud – but that still leaves the question about what to do with the remaining IT infrastructure and workloads that aren’t best suited in the cloud.
Just 12 months ago 10% of organisations had already shut down their enterprise and on-premises data centres and moved their infrastructure to colocation facilities. Gartner predicts that by 2025, 80% of enterprises will have moved to colocation.
Is this the right choice for you and your company? Read on.
On-prem – the good
There’s a lot to like about keeping your servers on-premises or inhouse. After all, they’re right there, if something goes wrong you can poke your head in and sort things out.
You also maintain complete control over your data, network and storage equipment. You make decisions about the data centre architecture, any expansion plans, upgrades and maintenance.
Put simply, you make the rules, you get to choose which people get access to what and what specialists you bring in when you need help!
Cooling, power, redundancy, bandwidth, and security measures are all monitored and managed by you.
Let’s face it – if your company is not planning to physically move anywhere anytime soon, keeping your servers where they are could seem like the easiest option…until you need to move, and it’s not.
On-prem - the bad
With great power comes great responsibility. When everything’s inhouse, all responsibility lands with you. If something goes wrong, the old metaphorical ‘one throat to choke’ will be you.
Building and running a data centre requires a lot of sunken cost and data centre maintenance and upgrades are an evergreen issue – data centres are never ‘set and forget’.
A data centre equally acts as an anchor for organisations, tying you to the premises it’s located, leaving little flexibility when it comes to rethinking office space and real-estate investments. With more than 50% of Australian workforces continuing to work from home permanently, does it make sense to continue keeping your IT equipment where your staff used to be?
You’ll need to buy ever-increasing amounts of WAN connectivity to connect your on-premises IT to cloud and other aaS providers in your IT stack. There are few companies nowadays that do it all in-house; with the majority using a hybrid mix of public cloud and other aaS IT solutions.
Additionally, when your IT is in one place and your cloud services are in another, your data is forced to travel, leaving you at the mercy of the highly variable performance of the public internet. Latency and jitter have a very real impact on the responsiveness of your applications, the performance of your team and the experience your users receive.
Your data centre and the building where it resides could very well face its own limitations around what it can provide. This acts as a major inhibitor to the ability to scale compute as quickly as the business needs. This in turn creates a barrier to your business achieving its primary goals and objectives within a timely manner.
Accompanying the rising amount of data is the increasing power and density of the systems that manage it. High density computing loads continue to grow rapidly, which poses a major risk for privately owned data centres.
As workloads and compute power grows, organisations are finding they can’t draw the power or provide the adequate environmental conditions and controls needed to support the specialist equipment housed in them.
To that point, if the data centre environment isn’t operating at maximum efficiency, your company will be bearing higher than necessary costs relating to power consumption and cooling. Power Usage Effectiveness (PUE) engineering and operating at the lowest PUE possible is crucial – it will be needed to protect you from downstream cost that you can’t plan for.
An interesting fact: privately operated data centres PUE usually sits at around 2.5. NEXTDC’s average PUE is 1.3, which directly hits your bottom line.
Sometimes small things are perceived as menial, but these ‘menial things can cause major issues. Server racks are locked using keys, and if that key is lost, or cannot be located quickly when it’s needed, “things” can really hit the fan.
On-prem - the ugly
Running your own on-prem data centre requires capital. And lots of it. When it’s not your core business, that money could be spent on growing the business.
There is so much that can go wrong in a data centre – it’s not just the servers you need to worry about, you have building maintenance, ageing mechanical systems, cooling, electrical and mechanical plants, UPS and more.
You foot the bill and the responsibility to hire the right staff to maintain the data centre environment and specialist infrastructure. If you have the head count – great – but finding readily available specialist resources that are skilled to build and run these ever-evolving facilities is increasingly difficult.
Skills shortages are one of the biggest problems cited by organisations globally. These aren’t general IT teams, they are specialised data centre technicians and engineers who understand the environment, infrastructure, and mechanical and electrical plants.
Operational best practice in the data centre is no longer negotiable. 80% of downtime occurs from human error therefore operational excellence is critical to ensuring near 0% human error. This takes significant training and investment of capital, resource and time and in some instances, third party verification of your practises being suited to the outcomes you’re chasing.
All operational costs sit with you – you can’t offload them, or the responsibility to someone else. The time and cost it takes to manage the data centre is self-owned. You may find yourself running to keep up with just keeping the data centre lights on, which has a major knock on effect to your growth objectives.
On top of the financial side of things, managing a data centre in-house means running the environment while at the same time developing and proactively executing the technology strategy to keep the business moving in the right direction.
Unless you’ve invested in a comprehensive Data Centre Infrastructure Management (DCIM) solution, you won’t have the sophisticated environmental monitoring and capacity planning tools needed to be successful.
Managing power usage and efficiency takes enormous effort – not just in designing the environment, but sourcing the technology and equipment used, and then maintaining it, tuning it, and testing it long-term.
If you’re a larger organisation with multiple data centres or premises, keeping them connected via fibre is enormously expensive (and doubly so if you decide to put in redundant fibre paths to protect against one link going down.)
The risk all sits on your shoulders – you are required to achieve and maintain ISO compliance (27001, 9001 etc). Equally, the risk of ensuring your DC environment is equipped to support and enable any business continuity or redundancy scenario is your responsibility.
Data centres are dangerous – they represent areas with high voltage electricity, working at heights and overhead work. They’re one of the most dangerous places for staff to work, so the safety practises in place must be rigorous to ensure the ongoing health and safety of your team.
Colo at NEXTDC
NEXTDC Colo facilities are staffed 24/7 – which is essential, because they manage the security, monitor the environment, tune it, carry out maintenance and so on. For each facility, it’s multiple people’s full-time jobs, and they work around the clock to make sure the critical infrastructure that keeps the facility humming along, is always running at its peak.
Many Colo providers offer a 99.999% uptime guarantee. According to the many online calculators, that represents 5-7 minutes of downtime per year. But the devil’s in the detail, and what they aren’t factoring in is the time it takes to cut over to a redundant environment if something fails. That can add up to hours per year when you take into account systems rebooting and reconfiguring to the primary or backup environment.
NEXTDC gives you a 100% uptime guarantee. No if’s or but’s – just a 100% uptime advantage.
When it comes to cloud, the proximity of your physical IT to your cloud access points is often overlooked. This is a critical consideration, you could have the fastest pipe in the world, but if your infrastructure is based on the other side of the country – physics dictates there will be delays. The good news is that this is avoidable.
When you host your infrastructure at NEXTDC, it means housing your IT in the same locations that host the major public cloud providers. NEXTDC’s national data centre footprint houses the country’s largest number of cloud onramps. This places your IT as close as it can physically get to the cloud - unlocking greater cloud capability and return on your cloud investments.
Instead of connecting over the internet, or private WAN links, you connect natively across the hall for just over $100 per month. When compared to the hundreds or thousands per month you’d be paying to connect your own data centre via a private fibre link, the savings are significant and are better suited being invested back into your business.
Remember how we mentioned earlier that losing the key to a server rack in your own on-premises data centre can be a difficult situation? NEXTDC’s data centres don’t use last century’s rack locking with keys. NEXTDC uses a rack locking system which can be unlocked via an ID access card, secured with your biometric fingerprint, or by authorised personnel via an intuitive customer management portal. No keys, no escorts, no time or money wasted.
The security measures at NEXTDC’s data centres are like something out of MI5. The cost to achieve this level of security could never be achieved by customers individually – but colo gives you immediate access to the best security in the world, at a fraction of the cost.
NEXTDC data centres are the only ones in the country with a five-star energy efficiency rating. A critical requirement to achieving sustainability objectives and keeping shareholders and customers happy.
Finally, NEXTDC take on all the risk in achieving the globally recognised certifications. They manage and pay for the audits, they maintain them and in turn, you get to leverage those at no cost.
NEXTDC: Helping your business grow
Quality is an absolute must when it comes to managing the critical infrastructure that drives your business. Outsourcing your IT to a colocation facility should be the easiest decision you make. That ‘one throat to choke’ is passed to your data centre provider – leaving you to focus on innovation and helping the business grow.
Decisions about the right way to go always come down to your specific environment, goals, and future plans. There is no one-size-fits-all solution, nor should it be approached that way. If you need help aligning your needs and priorities to a solution that will suit you now, and the years to come reach out to NEXTDC and request a complimentary strategy session.Contact NEXTDC