The hidden cost of cloud sprawl is similar to unaccounted costs associated with tasks in our everyday lives, like shopping for food and household items. In many cases, our shopping habits are inefficient, but we don’t take the time to analyze it thorough enough to take action.
We typically grocery shop at a variety of places. We have a convenience store around the corner usually at a higher price (managed services). We have choices from competing tiers of grocery store chains (public clouds) and an outlet-type store that requires bagging our own groceries (open source). The choice and variety are great for us as individuals and family shoppers. We can usually find what we need without relying on any one particular store (lock-in). Making a special trip to get something is not a hassle.
Total Cost of Ownership and Economies of Scale
Consider, however, that when buying groceries for significantly more people per household—or more consumers of IT—a little more planning goes a long way in terms of savings. If you would tally the per-person grocery bill when shopping for yourself or in a couple, you would most likely have a larger bill per person than shopping for five or six in a family. Include the cost for the time spent planning, coming and going, shopping in the store, and making decisions about what to buy. This is the overhead expense for shopping. No one typically thinks of themselves as an hourly employee when doing these domestic tasks, but the time adds up.
When we factor in the price for transportation, along with the time spent doing all of the associated tasks, we have a pretty compelling case for why it costs a lot more per person with fewer people to shop for in a household. It’s the same for IT. If you have people in your organization individually or even in small teams all going out to shop for IT resources from public and private cloud offerings, the overhead expense is factored in and you end up with cloud sprawl and hidden costs.
Essentially, when you add more people to the list you’re shopping for, trips could be planned in advance, less frequent, and the result will be that the added expense will be less per person. There’s a term—“Economies of Scale”—in business to describe this phenomenon. We can also add the term “Total Cost of Ownership” or (TCO) for this shopping scenario from end to end. It includes the planning and time spent going to and from a store.
Squelch Hidden Costs for Cloud Sprawl
What can you do to squelch the hidden cost of cloud sprawl? Adopt a centralized platform as a gateway to all your cloud resources to help mitigate the individual, duplicated effort. No one as an IT consumer will have to “shop around” and you could set up bulk ordering for any sets of resources common across several users and teams.
Taking it one step further, when enterprises grow and onboard franchises, satellite offices, or, in the case of banks, branch offices, they need a set of IT resources that are typically standardized. Theses are roughly the same for each store in a chain or whatever organizational entity exists at the front lines. You can see this in action when you go to the venues and see that the menus serving two chain restaurants are almost identical to each other no matter where you go. The digital services that run behind the scenes will likely be identical too.
Modern enterprises plan for the economies of scale associated with building out the IT infrastructure required to run these businesses. Instead of having each branch or store set up their own IT resources by business unit or create each one as a custom order, they configure them as “blueprints” where the resources are the same and they just need to spin them up with locally relevant configuration settings. The blueprint could grab the resources from the various “stores”, or public and private cloud resources available, and the “shopper” just needs to click Submit. The resources are then provisioned accordingly. A lot of time is saved by not having each new branch or store get these resources independently each time they want to add a new one to their portfolio.
Finally, once the blueprinted IT resources are in place, they can be updated regularly all at the same time if necessary. If a critical security issue is uncovered in a particular IT resource running in that environment, a patch could proactively be administered to the whole set to avoid any further problems.
CloudBolt provides exactly the type of platform that can help you achieve these economies of scale. You have one place to manage all the resources that IT consumers need as “shoppers.” You can do all of the upfront planning and significantly lower the TCO for each consumer-owner of IT resources.
For more information about advancing a cloud control strategy that not only helps mitigate hidden costs but also tackles security and user access, see Advancing a Cloud Control Strategy.
To see how CloudBolt can help you manage the hidden cost of sprawl request a demo today!
On Wednesday, May 22, SovLabs teamed up with Red Hat Ansible to offer a deep dive technical webinar, “Best Practices for Integrating vRealize Automation and Ansible Tower.” You can watch webinar in its entirety on the Red Hat Ansible website here and read our summary of the webinar here. Below is a summary of the brief Q&A session at the end of the webinar.
Car Buying—The Romance
Getting a new car from a dealer is often one of the most exciting times for any family or individual. Test driving, picking out colors and options, and then eventually driving off from the dealership with a brand new vehicle makes for an exhilarating event. You might want to show off your car, keep it clean, and start noticing others who have made similar choices. Some dealers even put a big wrapping present bow on your vehicle and do a photo shoot before you leave.
Car Ownership—The Reality
When reality sets in, though, it’s essential to follow the recommended service intervals set by the manufacturer and have records to keep your warranty in place, as improper care voids most warranties. A light comes on to remind you of services and you typically get a barrage of messages from the dealership, and sometimes recalls are issued to mitigate issues from car manufacturers when defective parts are discovered.
Likewise, for infrastructure provisioning in any private or public cloud environment, virtual machines (VMs) are issued as brand new, and, just like a car ,things need to be updated over time to run smoothly.
For a car, you have no choice but to manually schedule the vehicle for the maintenance and take the car “offline” so to speak to get the work done. That’s not the case with infrastructure—you can automate the whole process, and if you don’t you’re stuck with a dreaded manual process that inefficient and time-consuming.
Just recently, while getting down to work on something important on my laptop, I also needed a software update. Thinking the software update might be important too I lost thirty minutes waiting for the updates to complete. Not a big deal, because there’s plenty for me to do, but can you imagine if a similar scenario is playing out in a production environment for five hundred to tens of thousands of VMs?
Improving the efficiency of deploying and managing your VMs can have a huge impact when you’re scaling to large environments. Saving thirty minutes for any task would be huge savings for an enterprise. Most would argue that even saving five or ten minutes on a process that affects many end user is certainly worth it.
Having complete cloud control means that you’ve implemented a solution, like CloudBolt, to automate the process of keeping your infrastructure up to date. Instead of addressing a backlog of IT requested updates from end users, you can have a structure in place to automatically—or by an on-demand process—check for updates and then roll them out to any cloud-managed instance of those resources.
Two Stages
There are two stages in considering up-to-date infrastructure resources. The first stage is the initial provisioning that might need domain-specific configuration information and the proper networking information. You’ll also need to consider tagging for billing codes, and identifying business units or teams for the proper accounting of who’s spending what and where for chargeback/showback purposes. Everything that is done with the VM is easily tracked, like how the VIN number of your vehicle is a tag that can be used to figure out a lot of information about your car.
If the operating system you are deploying to the VM needs an update between major releases, you could build that into the orchestration process before it goes live to the end user. You could add security applications and specific agents for any kind of logging or your enterprise-specific monitoring tools. You’d also likely use a hostname template that automatically and correctly names the host and then can be used to help with any integrated solution in your enterprise for inventory and control of resources. You could apply that same VIN comparison here—perhaps an even tighter comparison than tagging in general.
The second stage occurs after the infrastructure is running and before decommissioning. There will be times you’d like to issue a patch to any software that you installed at runtime for the initial build of resources. You might need to add settings to a group of VMs and then restart services for the settings to take effect and you can do all of that remotely instead of a tedious manual process. As you manage the resources, you’ll have windows of time when the updates should occur and you can implement power schedules to only run the VMs when they’re needed and expiration dates for the decommissioning process.
This complete lifecycle control of your VMs from the initial provisioning, through the ongoing maintenance, and then retiring of the resources can be easily implemented with CloudBolt. The exhilaration of car buying is certainly not quite the same. But the dreaded manual process of car maintenance is a reminder that doing that for VMs does not scale well in the enterprise.
Ready to see how CloudBolt can help? Request a demo!