Skip to content
Managed Services · Infrastructure · Data Center

Data Center.

Design, implementation, assurance, operation and monitoring of your data center's technology infrastructure: compute, storage, network, virtualisation, backup and platform security, under a single line of technical accountability.

Book a diagnostic (60 min) Review the associated exposure
What it is
The availability of a platform is determined by its architecture. Operation can sustain it or degrade it, not increase it.

SUMāTO works across the five phases of the technology platform's life cycle under one criterion. The availability level is set as a business decision, with its cost made explicit; it is realised in the architecture and in the implementation; it is demonstrated through acceptance testing against criteria agreed beforehand; and it is sustained through documented procedures and telemetry. Scope is technological. Civil works and the building's electrical and mechanical infrastructure belong to the specialists in that discipline: the design hands them the power, heat rejection and space requirements, and coordination with them forms part of the service.

Typical situation

The platform grew through successive purchases, without a common architectural criterion.

The pattern recurs consistently: each project brought in its own equipment on its own criterion and through its own supplier. The cumulative effect is a platform whose aggregate behaviour under fault conditions is neither documented nor verified.

01

Capacity without projection

There is no projection of compute, storage and network identifying the point of exhaustion, so expansion is decided once the limit has already been reached.

02

Unverified redundancy

The architecture declares alternate paths and nodes that have never been exercised, so their behaviour on the loss of a component is an assumption.

03

Heterogeneous platform

Generations and vendors accumulated project by project raise the cost of operation and multiply the firmware and compatibility combinations to be sustained.

04

Outdated inventory

The absence of a current inventory and up-to-date diagrams prevents the impact of a change being assessed before it is authorised.

05

Recovery not exercised

Backups exist, but restoration is not tested regularly, so the recovery time is an estimate.

06

Operation without telemetry

Deviations are detected by report rather than by threshold, which removes the window for preventive intervention.

Scope

Five phases across the technology platform.

Scope is contracted in full or by phase. In either case the architectural criterion and the acceptance criteria are the same, and traceability between phases is preserved.

01

Architecture design

Target availability level and architecture for compute, storage, network and virtualisation, with the investment attached to each option.

02

Sizing and capacity

Load projection to an agreed horizon and density per cabinet, with the power, heat rejection and space requirements handed to the physical infrastructure specialist.

03

Specification and selection

Technical specification fit for tender, and comparison of alternatives against the requirement and their total cost of operation.

04

Implementation and integration

Deployment and configuration of the equipment, integration with the existing platform and with the services that depend on it.

05

Workload migration

Migration sequence, agreed windows and a rollback procedure defined before each workload is touched.

06

Assurance and testing

Component and integrated testing, including failover and restoration from backup, against criteria defined before execution.

07

Operation

Platform administration, patch and firmware management, capacity management, change control and documented procedures.

08

Monitoring

Telemetry of availability, performance and capacity consumption, with defined thresholds, alert handling and historical series.

Exposure

Exposure associated with a platform that has no single criterion.

These risks are neither simultaneous nor immediate. They accumulate while the platform operates within its margins, and they materialise when an event reduces those margins.

01

Single point of failure

A component with no effective alternate path turns any planned maintenance into a window of service unavailability.

02

Early capacity exhaustion

Growth stops on available capacity sooner than expected, placing the expansion outside the budget cycle.

03

Obsolescence with no replacement path

Equipment out of vendor support carries production services, with no replacement route and no budget assigned.

04

Inherited exposure

Configurations and access inherited from earlier projects remain active because nobody holds the criterion to withdraw them.

05

Substantiation under audit

The absence of a current inventory, test protocols and a change log prevents compliance being evidenced to a third party.

06

Continuity not demonstrated

The recovery plan has not been exercised under real load, so its recovery time is an estimate rather than a measurement.

Outcomes

Outcomes of managing the full cycle under one criterion.

Availability determined by design

The target level is set at the outset with its associated cost, and verified by testing before entry into production.

Capacity with a defined horizon

Compute, storage and network projected over time, which moves expansion into the budget planning cycle.

A platform on one criterion

Purchasing decisions answer to a reference architecture, which lowers the cost of operating disparate generations and vendors.

Accountability not fragmented

One technical owner covers design, implementation, testing and operation, which removes the grey area between suppliers.

Documentary evidence available

A current inventory, signed protocols and a change log remain available as substantiation for internal or external audit.

Decisions on historical series

Expansion, renewal and deferral decisions rest on measured trend rather than on judgement alone.

Service model

Structure of the service.

Execution is by phase, with a control point between each. A single phase can be contracted in isolation, provided the preceding one is received documented.

01

Phases with a control point

Each phase requires formal acceptance of the preceding one, with its deliverables and criteria met, before it begins.

02

Reference architecture

Sizing rests on a documented reference architecture and on the practices of the selected vendor.

03

Documented scope

Scope, exclusions and design assumptions are documented before any contractual commitment.

04

Acceptance criteria set in advance

Criteria are agreed before the test is executed, and acceptance is granted against the recorded result.

05

Service levels by criticality

Response and resolution times differentiated according to the impact of the event on the operation.

06

Periodic reporting

Availability, capacity consumption, changes executed and maintenance, with the series that supports the next decision.

The SUMāTO approach

Independent technical criterion and documentary evidence.

The operating phase draws on the capabilities that sustain our managed services, and can integrate with NOC, SOC and Managed Services under common governance.

01

Vendor independence

The requirement is established in the design; brand selection is resolved afterwards, comparing alternatives against that requirement and their total cost of operation.

02

Design precedes specification

Availability level and load projection precede equipment selection, and not the other way round.

03

Assurance with evidence

Every test leaves a protocol, a result and a named owner. What has not been tested is declared outstanding, explicitly.

04

Continuity from design to operation

The operating team knows the design assumptions, so a deviation is classified as such rather than as a normal condition.

05

Governance and service levels

Measurable commitments, defined roles and periodic reporting as the basis for accountability to the business.

06

Improvement on root cause

Every incident closes with a root cause and with the architecture, procedure or threshold change that prevents recurrence.

Deliverables and SLA

Deliverables by phase and service commitments.

  • Architecture document with target availability level, assumptions and load projection.
  • Power, heat rejection and space requirements for the physical infrastructure specialist.
  • Technical specification for tender, with criteria for comparing alternatives.
  • Implementation and migration plan, with windows, acceptance criteria and rollback procedure.
  • Test protocols and acceptance records by component and for the integrated tests.
  • Current inventory, up-to-date diagrams and the platform change log.
  • Operating and recovery procedures, with a maintenance and update plan.
  • Service Level Agreement (SLA) with response and resolution times by criticality.
  • Periodic operations report: availability, remaining capacity, changes and trend.
Frequently asked questions

Frequently asked questions on scope.

Does it include civil works, power or cooling for the building?+
No. Scope is technological: compute, storage, network, virtualisation, backup and platform security. The building's electrical and mechanical infrastructure belongs to the specialists in that discipline. What the design does deliver is the power, heat rejection and space requirements those specialists need, and coordination with them forms part of the service.
Is it necessary to contract all five phases?+
No. A single phase can be contracted in isolation. The condition is receiving the preceding phase documented: operating a platform with no architecture and no test protocols means assuming conditions that were never recorded.
Does it apply to an existing platform?+
That is the most common case. Work begins with a survey of the real state —inventory, effective redundancy, remaining capacity and current vendor support— and that survey determines the sequence of intervention.
How is the availability level determined?+
It derives from the economic impact of unavailability and the processes affected, not from the platform. That figure justifies the level of redundancy, and the design presents the capital and operating cost of each option.
Can it be executed without interrupting operations?+
In most cases yes, through staged migration, agreed windows and a rollback procedure prepared before intervening. The sequence is defined in the design and each window is approved individually.
Does the scope include cloud platforms?+
Scope covers owned infrastructure. Defining which workloads are best moved to cloud platforms is addressed in Cloud, on the same availability and cost criteria.
Does it replace the internal infrastructure team?+
It can take on the full cycle or intervene in the phases requiring specialist depth —architecture, acceptance testing or capacity management— with day-to-day operation retained by the internal team.
Starting point

IT Maturity Assessment

Any intervention on an existing platform starts from a baseline. The assessment inventories the infrastructure, evaluates maturity by domain and establishes the sequence of intervention on a priority basis.

First step

Establish the baseline for your platform.

In a 60-minute session we review current capacity against projected, effective redundancy on the loss of a component, and equipment out of vendor support, with a preliminary indication of priorities.

Book a diagnostic (60 min)