Techbanq

Cybersecurity · architecture

The strongest controls are drawn before the system exists

A structural security decision — where the trust boundaries sit, which component may talk to which, where the sensitive data is allowed to travel — costs a drawing change at design time and a re-platforming after launch. Architecture work moves those decisions to the point where they are still cheap, and makes whole classes of attack impossible rather than merely detected.

Cybersecurity credentials

  • National Cybersecurity Authority, Saudi Arabia

    NCA registration

    Registered cybersecurity service provider

  • ISO

    ISO/IEC 27001

    ISO/IEC 27001 certified

Certifications & registrations

Design before defence

A penetration test finds the flaw; the architecture decided it could exist

We do not have an office in Saudi Arabia

We serve organisations there from our Riyadh office and our engineering hubs in Lahore and Karachi, travelling for the parts of an engagement that need to happen in person.

When a tester walks from a compromised web server to the customer database, the finding is the walk — but the cause is a design in which that path existed at all. Testing after build answers "what can an attacker do to this system as it stands"; architecture work answers the prior question, "what should this system be structurally incapable of allowing". A flat internal network, a service holding credentials it only needs once a year, sensitive data flowing through components that never needed to see it: none of these are bugs a patch fixes. They are decisions, and the only cheap moment to make them differently is before the system exists.

The same discipline applied to a live estate is a review rather than a design, and it is worth doing precisely because most estates were never designed — they accreted, acquisition by acquisition and project by project. The review reads the estate as if it were a proposal: where does trust actually flow, which boundaries exist on the diagram but not in the configuration, which single component would take the most down with it. The output is not a list of missing patches; it is the short list of structural changes that would remove entire categories of finding from every future test.

Two different questions

What a design review answers that a penetration test never will

The two are complements, not rivals — but they are bought for different reasons, at different moments, and confusing them buys the wrong one.

OptionWhen in the system's life it runsWhat it examinesWhat acting on a finding costsWhat it provesBest for
Security architecture reviewAt design time, or on a live estate before the next major change.The structure: trust boundaries, data flows, blast radii, the controls each threat is met by.A drawing change before build; a planned structural project after.That whole classes of attack have no path, not merely no known instance.Before building, before a major change, or when tests keep finding the same class of flaw.
Penetration testAfter build, against the running system.The implementation: what an attacker can actually reach and chain today.A fix per finding — a patch, a configuration, occasionally a redesign the review would have caught.That the build as deployed resists the attacks that were tried against it.Validating a built system, meeting an assurance requirement, testing the estate as attackers meet it.

The method

How a design is examined before a line of it runs

STAGE 01

Assemble the design evidence

Diagrams, data flows and trust boundaries where documents exist; structured workshops with the architects where they do not. Undocumented estates are the norm, not a blocker. One picture of the system both sides agree is the real one.

STAGE 02

Model the threats against the paths

For each asset worth stealing or stopping, the plausible routes to it are walked on paper — through the interfaces, the trust relationships and the failure modes. A threat model tied to this design, not a generic checklist.

STAGE 03

Judge each control against its threat

Every control earns its place by the threat it meets; every threat without a control is named. Controls that exist for no modelled threat are candidates for removal — complexity is also a risk. Gaps and redundancies, each traced to a threat or its absence.

STAGE 04

Order the findings by structural payoff

The recommendations are ranked by how many future findings each one removes — the boundary that shrinks the blast radius outranks ten point fixes. A short list your roadmap can actually absorb, with reasons.

Our role

What we challenge in the design and what your architects own

  1. 01

    An outside adversarial reading

    We read the design the way an attacker would, without the familiarity that makes a team stop seeing its own assumptions — the value of the review is exactly that distance.

  2. 02

    Decisions made explicit

    Every trust relationship and every data flow is written down as a decision someone can defend or change — because the dangerous ones are the ones nobody remembers making.

  3. 03

    Reference patterns, not bespoke exotica

    Where a standard pattern solves the problem — a broker between zones, a token exchange, a signing boundary — we recommend the pattern your team can find documentation for, not an invention only we understand.

The trade-offs — what the business will pay in delivery time for which reduction in structural risk — are decisions your architects and owners make; we make sure they are made knowingly, with the paths and the costs on the table.

Common questions

What buyers ask

Is an architecture review still worth it once the system is already in production?+

Yes, and for most organisations that is when it happens, because most estates were accreted rather than designed. On a live estate the review reads the structure as built — where trust actually flows, which boundaries exist on paper but not in configuration — and its findings become the structural items on the roadmap: the segmentation project, the credential broker, the data flow that should be cut. The moment it pays best is just before a major change, when the structure is about to be touched anyway and a better shape costs least to adopt.

What can you review if our architecture was never documented?+

The system itself, through the people who run it. Undocumented estates are the normal case, so the review begins with structured workshops that draw the architecture out of the architects and operators — the components, the flows, the trust relationships as they actually exist. That reconstruction is itself a deliverable: the first accurate picture of the estate many organisations have had, and the map every later decision gets made against. What we do not do is decline the review until documents exist, because the estates most in need of review are precisely the undocumented ones.

How does threat modelling stay useful after the workshop ends?+

By being tied to the design rather than the date. A threat model built against your actual components and flows stays valid until those change, and it becomes the standing checklist for change: a new integration, a new data flow or a new external dependency is checked against the model in a design discussion rather than a fresh engagement. We hand the model over in a form your team can maintain — and the sign it is working is that it gets amended in your architecture reviews, not re-commissioned from us every year.

Next step

Put a design in front of us

A review of one design — a new build or a live estate — that returns the structural risks ranked by what each costs to carry versus to fix.