istimthalOriginal model example

A guided company showcase

Change the constraints.
See what holds.

Explore prepared scenarios. Turn up demand, reserve capacity and adjust the search effort. The model recalculates the result each time.

The original question

Fit the same workloads into a less expensive fleet without exceeding its capacity.

From the Istimthal model example

This hosted showcase uses synthetic Istimthal scenarios and returns checked summaries. Your company’s data cannot be submitted here. Custom problems and deliverable solutions belong in an agreed commercial pilot.

Your presentation allowance

Paste your invitation link or code to unlock the showcase.

An invitation permits up to 10 searches and expires within 14 days. Its allowance is enforced on the server and shared across devices. Reloading or clearing cookies does not renew it.

Held only while this page is open. After reloading, reopen your invitation link or paste it again. No access cookie or browser storage.

Choose the challenge

Loading the challenge

— workloads

The problem statement will appear here.

Adjust the constraints and search effort

More effort explores more candidates within this bounded model.

An invitation is required. Paste your invitation link or code to unlock your presentation allowance.

Loading the prepared scenarios. Adjust the controls before using a run.

An accepted search consumes one run, including a tie, no-go or interrupted request. Loading scenarios and checking access use no runs. The showcase returns aggregate results, without workload assignments or downloadable solutions.

Read the result

Checked summary

Run the selected scenario to compare basic first-fit, five fixed heuristic rules and the searched candidate under the same constraints.

Five-rule heuristic—example cost units
Candidate cost—example cost units
Beyond the heuristic—against the five-rule heuristic
Basic first-fit— cost units— hosts
Five rules— hosts
Candidate— hosts

The basic first-fit comparison will also appear here.

Awaiting your search

Every workload assigned exactly once
—
CPU and memory within capacity
—
Selected headroom policy
—
Actual search effort
—

A separate server checker reconstructs assignments, capacity and cost. All three allocations use the same selected capacity policy.

Read the 72-case benchmark and methods.

These are aggregate results for the chosen synthetic scenario. Detailed allocations, company-data evaluation and implementation deliverables are reserved for an agreed commercial pilot.

What this showcase demonstrates

Istimthal searches real candidate allocations within the effort you select. Demand and headroom change the calculation. The checker verifies all three allocations independently against the same requirements. The stronger reference takes the best of five fixed ordering and packing rules, including cost-aware opening. The search explores more rules within your selected budget. Its headline reduction uses that stronger reference; the basic first-fit comparison remains visible.

A scenario may improve, tie or fail to beat a reference. The five-rule portfolio is a simple heuristic comparison, not a benchmark against a commercial solver. This bounded search does not establish a global optimum, prove production reliability or promise real-world savings. It is the hosted synthetic allocation model from Istimthal’s web example, rather than the full Istimthal engine.

See the method.
Scope your own pilot.

The memory-pressure challenge combines 48 workloads with competing CPU and memory needs across five host types. Raising demand by 20% while reserving 20% capacity makes cheap, individually attractive host choices harder to combine into a good fleet.

Change the scenario or constraints and run again. The result comes from a new search, not a stored answer. The summary shows the actual effort, costs and checks while keeping the detailed allocation on the server.

The underlying model originates from Istimthal. A company-specific problem needs agreed objectives, data handling, evaluation criteria and commercial terms before we solve it.

Discuss a commercial pilot