SCNET · Enterprise IT · Ankara, Türkiye

Sanal Çekirdek

Decide where each workload belongs — on evidence.

Colocation means running your own hardware in a third-party data center facility. Sanal Çekirdek never evaluates it in isolation: it is compared against on-premises infrastructure and suitable public cloud options within one placement decision.

Placement is the most expensive decision to reverse. When performance, data class, licensing model and dependencies are not measured up front, the cost does not surface in the first year — it surfaces in the third.

What decides placement

Not every workload belongs in the same place. Latency sensitivity, data class and legal location requirements, how the licensing model treats virtualization and core counts, integration dependencies and exit cost are weighted in one table. The decision follows criteria, not preference.

  • Latency, throughput and predictability requirements
  • Data class, location and access constraints
  • Licensing model: core, socket and virtualization effects
  • Dependencies and cost of exit

When colocation is the right answer

Hardware still under depreciation, workloads that need dedicated accelerators, cases where data location is fixed by contract, and steady loads whose unit economics are unpredictable in public cloud all favor colocation. The same reasoning does not hold for variable or seasonal demand.

  • Existing hardware still under depreciation
  • Dedicated hardware and accelerator requirements
  • Data location fixed by contract
  • Steady, predictable load profile

Hardware and software sourcing

Sourcing follows the architecture decision, never precedes it. The requirement list is derived from the target architecture, the licensing model is matched to actual usage, and warranty and support scope is chosen against workload criticality. We work vendor-neutral and document why each choice was made.

  • Requirement list derived from target architecture
  • Licensing model matched to actual usage
  • Warranty and support scope set by criticality
  • Lifecycle and refresh schedule defined up front

Migration and operational handover

A migration is a project with acceptance criteria, not an overnight task. Inventory and dependency mapping come first, rack and cabling plans follow, and the migration window is reconciled with the business calendar. Handover completes when the acceptance criteria are met.

  • Inventory and dependency map
  • Rack, power and cabling plan
  • Migration window and rollback plan
  • Acceptance criteria and operational handover

How we work

  1. Map the workload inventory and its dependencies
  2. Weight the placement criteria together
  3. Compare options within one cost model
  4. Write the migration plan and acceptance criteria
  5. Complete handover and review the decision on a cycle

How success is measured

  • The placement decision is documented and traceable
  • The planned migration window was not exceeded
  • Licensing and support scope match the inventory exactly
  • Unit economics are tracked against the baseline

Frequently asked questions

Colocation or cloud?

The question is answered per workload, not per company. Within one organization some workloads run correctly in colocation and others in cloud; the criteria above are weighted and the reasoning is written down.

Can we move our existing hardware?

Usually yes. Moving hardware still under depreciation is often the lowest-cost option while warranty and support remain in force; age and refresh schedule are assessed together during the inventory work.

Which vendors do you source from?

We work vendor-neutral. Selection follows the requirements of the target architecture and the licensing model, and the comparison table showing why a product was chosen is delivered with it.

If you would rather decide placement on criteria than on instinct, share your workload list and we will build the comparison table together.

Request a placement comparison