You can put just about everything in the cloud these days: your website, customer data, marketing analytics, employee payroll, human resources, email and messaging and darn near anything else that you can think of. A reliable cloud service provider will take care of all your storage and/or hosting needs. In fact, the cloud has become so versatile and ubiquitous that it’s putting the kibosh on traditional IT. Well, maybe not completely – not just yet, anyways – but it is nudging traditional information technology in a new direction.
Historically, IT has been, let’s say, somewhat less than transparent when it comes to its costs. In fact, it’s often been downright opaque. Most businesses know what they’re paying for old school IT, but they don’t know why they’re paying it; they don’t know how the costs are broken down. It’s not readily apparent what the cost is to deploy a particular in-house data center service, because individual costs are lost in a seas of jargon-ey talk about different initiatives or conveniently hidden among one massive project-level budget.
Things are much more straightforward with the cloud. You’re not buying any physical property, nor are you building any complicated infrastructure. Your cloud provider already has everything you need in place at its data center. Costs aren’t hidden, and cloud services have an unbelievably attractive value proposition. You have the option of paying only for what you need, scaling up or down as needed and receiving incredible IT service that isn’t afraid of transitioning to new technologies to meet changing business necessities.
Choosing cloud over traditional methodology will save you big time, but it can save you even bigger time – let’s go ahead and say “huge time” – if you take a few steps to control your costs. There are several different things that you can do to mitigate cloud cost inconsistencies and allow yourself to accurately predict and budget for cloud costs. Let’s explore them, shall we?
Use a Cloud Management System
Want to keep the cloud under control? Consider implementing a cloud management system capable of controlling the decommissioning and provisioning of your cloud instances as well as perpetually tracking your expenditures on it and measuring them against your planned budget.
Dell Enstratius, RightScale and Scalr are just a few of the options you have to choose from on this front, and they all have both on-premises and software-as-a-service models to best suit your needs. In addition, workflows, support for advanced monitoring triggers and policy automation are some of the other benefits you’ll get from implementing one of the above tools. Coming up with a plan to manage your cloud is great, but consistently following through on it in the long-term will be an extreme challenge without a management system in place.
You may think you have the time and ability to do it all manually, and maybe you do…for a little while. But don’t make the mistake of thinking you’re always going to be able to track every single thing happening in your cloud instance. Sooner or later you’re going to get busy with other business needs. If you want to ensure your costs stay where you want them to be, then get the help you need to do so.
Migrate as Much as You Can to the Cloud
What can you migrate to the cloud? Quite a bit, as we discussed briefly at the beginning of this blog post. Big Data computations, DNS and application messaging queues, load balancing and relational databases are all examples of things that can go in the cloud.
You can save oodles of cash – and who amongst us couldn’t use a few more cash oodles? – by replacing a convoluted mirrored database design with a simplified one in the cloud. How? You’ll do away with multiple compute instances, multiple operating systems, multiple database licenses and complex mirroring configurations, backups, monitoring and management. Not only will you save on all of those costs and the operational expenses associated with not configuring systematically through an online interface, but you can reduce the complexity of the whole thing by as much as 50 percent.
Size up Your Instances
With the elements you want to shift to the cloud earmarked for the big move, it’s time to figure out what size servers are needed. Conducting a “right-sizing” exercise will help you to figure out what the best size instance is for moving your exiting servers over. Good news: this stuff is pretty easy. All you need to do is watch over your current servers and track utilization for a pre-determined period of time, then rely on the steady-state average utilization to come to a cloud instance size conclusion. If you need one, there are tools available for helping you to capture this type of info.
Knowing what your CPU, disk and memory usage stats are over a timeline of one to three months will help you figure out your steady-state utilization. It’s important to right-size and avoid under or overshooting. Yes, you can always scale later, but you want to have the correct starting point.
Establish Workload Classifications
With your server sizing completed, it’s then time to move onto figuring out whether individual items in your cloud instance are mission-critical and therefore should always be on only need to be run occasionally.
If a server is going to be utilized for less than a majority of a month or less than half of the year then it’s probably best suited for traditional, non-committal cloud instances.
For something that’s going to be running the entire month, reserved or dedicated cloud instances are probably the way to go in terms of being cost-responsible. These types of instances are ones that are agreed to be run for a period of at least a year. You’re likely to pay more upfront for this reservation, but you’ll save in the long run on the hourly costs.
With autoscaling in place your instances will be adjusted based upon load or performance triggers. This can be complicated, but it can also save you major moolah. There are some key metrics you’ll need along with excellent knowledge and insight of your application: what your triggers to increase or decrease your storage space will be; how many transactions or sessions your application can run with minimal configuration; the amount of additional transactions or sessions it could run were you to add an online or app server; and the base configuration for hitting your steady-state user load.
Autoscaling allows you to run things with a lower number of instances the vast majority of the time when it’s properly implemented, and those third-party toolset examples we mentioned earlier have it built right in. You’ll end up running only the amount of cloud instances necessary at any given time rather than those needed to handle those rare peak loads.
One of the biggest reasons organizations make the move to the cloud is to save money. By following these steps you can further enhance your savings. But no matter what you decide to do, you’re going to save on IT costs by shifting from traditional to cloud. So why are you waiting to make the move? Do it now and save!
Image Source: PC Tech Mag
Find out more about Nick Santangelo on Google Plus