Seshail KamannaDigital Transformation
← The eight practice areas Stage 02 · Review system and architecture

Review system and architecture

An estate is rarely bad by design. It becomes bad by accumulation — each decision defensible on the day it was taken, the sum of them impossible to change.

Key offerings

System review · System architecture review · Network architecture review

Identify improvement in architecting components (services, data, app, compute, storage and network) to enable most optimal landscape which helps in composability, open standards, in enabling speed of change

The question is not what you have, it is what it costs you to change

Most architecture reviews produce an inventory. An inventory tells you what exists; it does not tell you why a straightforward product change takes nine months, why the same customer record is authoritative in four places, or why the annual run cost rises faster than the business it supports.

So the review is framed around change velocity rather than completeness. Across services, data, applications, compute, storage and network, the test is the same: does this landscape let the organisation move at the speed its market demands, and if not, precisely which components are holding it?

Composability and open standards

The estates that age well share two properties. Capabilities are composable — assembled from parts that can be replaced individually rather than a monolith that must be replaced entirely — and the interfaces between them follow open standards rather than a single vendor's conventions. Neither is an ideological preference. Both are simply what makes the next decade of change affordable.

That said, a target architecture the organisation cannot staff or fund is not a target, it is a wish. The recommendation has to survive the client's actual capability and cash profile, which is why remediation is sequenced rather than presented whole.

Scope

What the review covers.

Application portfolio

Every application against business capability, cost, condition and criticality — identifying overlap, orphans and obsolescence.

Data architecture

Where each entity is mastered, how it flows, where it is duplicated, and what that duplication costs in reconciliation effort.

Integration and interfaces

Point-to-point sprawl against managed integration, and whether interfaces can be changed without coordinated release.

Compute and storage

Placement, utilisation and sizing across on-premise, colocation and cloud, with the actual economics rather than the list price.

Network architecture

Topology, resilience, latency to the places work happens, and whether the design still matches how the organisation operates.

Security posture

Architectural exposure — segmentation, identity, privileged access and the assumptions the design quietly depends on.

Technical debt and obsolescence

What is unsupported, what is nearly unsupported, and the deadline each imposes whether or not it is acknowledged.

Non-functional performance

Availability, recovery, scalability and performance measured against what the business genuinely requires, not what was specified.

Target state and gap

A defensible destination, the distance to it, and the order in which the distance should be closed.

Method

How a review runs.

Six to ten weeks for a mid-sized estate. The output is a decision paper, not a catalogue.

  1. Frame the question

    What decision the review must inform — an investment, a migration, an acquisition, a stalled programme. A review without a decision attached is rarely acted upon.

  2. Assemble the picture

    Applications, data flows, infrastructure, contracts and cost, drawn from records and interviews and reconciled where the two disagree.

  3. Test against principles

    Composability, open standards, resilience, security and cost-to-change — applied as tests with evidence, not as adjectives.

  4. Find the real constraints

    Which components genuinely limit change, distinguished from those that are merely old. Age is not a defect; inflexibility is.

  5. Define the target

    A target state sized to the organisation's funding and capability, with the trade-offs of each option stated plainly.

  6. Sequence the remediation

    Ordered by dependency, risk and payback, with the first phase specified closely enough to start.

Selected work

Architecture and estate reviews.

Clients are described by sector rather than named.

The full sequence

The other seven stages.

Want to know what your estate actually costs you in speed?

Get in touch