pixstock - Fotolia
IT is expensive and -- worse yet -- a lot of that expense isn't even necessary.
An average organization will spend around 70% to 80% of the IT budget to keep the lights on: run the IT platform with zero additional value to the business itself. IT cost reduction problems persist alongside the advances of DevOps, distributed cloud-native applications and containers.
IT budget planning is difficult without a baseline for comparison. Few organizations baseline IT costs comprehensively. For example, energy used by cooling and auxiliary power systems in the data center may be paid for by facilities management, while maintenance and support costs for servers, storage and network systems come out of the IT budget.
Even so, basic IT cost reduction can arise out of basic information. For example, what are the resource utilization rates across all the servers? Can virtualization lower the number of servers in use, cutting down hardware, maintenance, license and power costs? The knock-on effect of higher utilization rates and less equipment is that less cooling and auxiliary power is required. Most organizations run at a level where $1 of IT equipment power costs requires at least $1 of auxiliary systems power costs. Get IT and facilities to work together on budget planning to create immediate and considerable savings.
Even the move to modern IT hosted on virtualization and cloud computing, backed up by continuous delivery processes rooted in DevOps, has not improved IT cost reduction.
Cloud computing has become a relatively user-friendly IT platform. The IT organization hands end users self-service capabilities so they can create a new instance of a service or application as required. Users and IT reap benefits, but it can also wreak havoc on IT spending.
Each image spun up in the cloud consumes basic resources, which are not free. However, the biggest costs in self-service clouds are zombie images: systems that are spun up, used for a while and then discarded. Rather than elegantly turning off the instance, the user might forget about it, leaving the resources and services it consumes idle and useless. A zombie VM, in most cases, has a full, licensed operating system -- along with its associated maintenance contract. The rest of the zombie stack -- application server, application and database -- continues merrily spinning around on cloud infrastructure, costing the company without adding any value.
The shortcomings of IT budget planning are nowhere more clear than within DevOps. Developers need test beds; self-service capabilities to spin up temporary systems are the perfect solution. However, it is too much of a bother for a developer to deconstruct the VM afterwards -- far easier to just leave it there.
Minimize IT licensing costs with a move to container technologies, such as Docker or LXD Linux containers. Containers share the OS, so there's no added OS cost when containers sit idle. This is just a scratch on the surface, though.
A full IT lifecycle management system logs VM and container usage and raises an alert when a system rests unused for a defined period of time. Some systems will automatically spin down the machine and recover all its licenses.
Developers like to choose their own tools, downloading them from sites as they see fit -- what's important to the developer is that a tool enables them to work to the best of their capabilities. However, the whole development team is then compromised by a lack of collaboration. The processes behind successful DevOps cannot be brought to bear without streamlined workflows across common tools. Do not attempt to force developers to use one set of basic tools, but do ensure that the right overarching tools pull everything together as needed. Tools such as Automic, Cloudify and Electric Cloud then enable greater process automation through to operations in production, and also build in feedback loops for developers and app owners.
Automation in its many guises is a major key to IT cost reduction. The vast majority of problems in an IT platform are caused by the carbon-based life forms who run it, rarely by the silicon-based components.
IT automation systems that apply patches and updates, provision systems and monitor and manage production contribute to cost control and prevent latent technical problems from escalating into expensive business issues. Automation allows workloads to move consistently across the hybrid IT platform. It allows IT shops to more easily and confidently migrate workloads when a new platform will serve the business needs -- and IT budget -- better.
A cloud-hosted model may not reduce IT costs, but it will eliminate hidden, unpredictable wastes of money. All those times sys admins either twiddle their thumbs waiting for something to happen or dash around in a frenzy when something has happened roll up into one repeatable operating expense each month. Cloud providers share the cost of administration across multiple customers. Cloud-scale operations enable IT cost reductions by spreading the cost of provisioning new IT and other data center hardware, maintenance and licensing, as well as other responsibilities among multiple customers.
Regard cost savings around the IT platform as a journey. Focus on the most easily achieved IT cost reductions first, remediate issues with automation and replace capital expenses with operational ones in longer-term IT budget planning.
Cloud costs can skyrocket when you stop looking; step up management
Move to the cloud -- now find an advisor for cost control and app monitoring
Agile and DevOps are expensive -- but there are ways to reduce the debt