Data Center and Cloud | DevOps Peach

Compare infrastructure operations in data centers, cloud, and hybrid environments.

Operations is more than keeping servers alive

System Operations (SysOps) covers creating and configuring networking, compute, storage, application deployment, backup, monitoring, security, and system changes. A data center requires CapEx and capacity bought in advance; cloud supplies resources on demand but does not guarantee every workload is always cheaper.

flowchart LR
  T["IaC template<br/>CloudFormation, Terraform"] --> D["DEV"]
  T --> E["TEST"]
  T --> P["PRODUCTION"]
  P --> M["Monitor<br/>CloudWatch"]
  M --> C["Change<br/>patch, scale, retire"]
  C --> T
  %% step T: One template describes the whole environment, so it can be reviewed and versioned.
  %% step D: The same template builds a development copy in minutes.
  %% step E: A test environment that matches production catches problems earlier.
  %% step P: Production is built the same way, not by hand.
  %% step M: Metrics and logs show how the system behaves under real load.
  %% step C: Changes go back into the template, so every environment stays in step.

Data center, cloud, hybrid

	On-premises data center	Public cloud	Hybrid
Spending	CapEx: buy hardware up front	OpEx: pay for what you use	Both
Capacity	Plan months ahead, often over-buy	Add or remove in minutes	Burst to cloud when needed
Who runs the hardware	Your team: power, cooling, racks	The provider	Split
Typical reason	Data rules, existing investment	Speed, global reach, elasticity	Gradual migration, latency to on-prem systems

In the cloud the shared responsibility model applies: AWS secures the physical facilities, hardware and the virtualization layer ("security of the cloud"); you secure what you put on it — operating systems, network rules, identities, data and encryption ("security in the cloud"). The split moves with the service: on EC2 you patch the OS, on a managed service such as RDS AWS patches the database engine.

Infrastructure as Code

Infrastructure as Code (IaC) makes environments self-describing, repeatable, and less dependent on manual configuration. On AWS the native tool is CloudFormation (templates in YAML or JSON, deployed as stacks); Terraform and the AWS CDK are common alternatives. Benefits:

Repeatability — dev, test and production are built from the same source, so "works in test, fails in prod" happens less.
Review — a change is a pull request that teammates read before it reaches AWS.
Recovery — if a Region or account is lost, you can rebuild from code instead of memory.

Hybrid cloud connects on-premises resources to cloud resources when a complete migration is unsuitable, for example over Site-to-Site VPN or AWS Direct Connect.

Create, Deploy, Monitor, Change

The SysOps lifecycle has four recurring activities:

Create — provision instances, databases, storage and routing, ideally from templates.
Deploy — build environments or copy them to another Region for disaster recovery.
Monitor — collect metrics and logs, alarm on problems, and scale for load.
Change — patch, resize, retire resources and release new versions safely, for example with blue/green deployments where traffic switches to a new copy only after it is healthy.

Cost is an operations job

Cloud cost follows usage, so an idle instance or forgotten volume still costs money. Tag resources with owner and environment, review spend in Cost Explorer, and set AWS Budgets alerts. Stop or delete lab resources when you finish — a habit that matters from your first day.

Command	Purpose
`aws cloudformation list-stacks`	View IaC stacks
`aws ec2 describe-instances`	Inventory instances
`aws ce get-cost-and-usage ...`	Query spend with Cost Explorer

Traps: assuming cloud is always cheaper; thinking AWS patches your EC2 operating system; making production changes by hand that no template records.