Skip navigation, go to main content
Contact

Securing territory

Control what decides the outcome; you need not own everything

A Go player does not win by capturing every enemy stone, but by enclosing more territory.

The same holds for digital infrastructure. What matters is not owning every server, but where the data sits, who can operate it, and what you lose if you change supplier. We build systems on open source hosted in Vietnam, then hand back full control.

The same nine points of territory: six stones enclose it in the corner, while eight in the centre still leave the wall open
Principle

Every piece of paid software has an open source equivalent

The difference is not knowing the names of these platforms. It is having someone operate them to a commitment. That is what we sell.

What you pay for now Alternative platform The difference
Power BI · Tableau Apache Superset No per-user licence fee
Google Workspace · Microsoft 365 iRedMail Data held in Viet Nam
Oracle Database PostgreSQL Migration service available
Redis (licence model changed) Dragonfly Protocol compatible
Confluent Kafka Redpanda Lighter to operate
Foreign cloud storage S3 operated in Viet Nam Data sovereignty

Migration is not always worth doing. A system that runs stably, is near the end of its life, or is tied to business software that runs only on a commercial platform should be left as it is. We operate both sides, so we can judge when a migration would not yet pay off.

Services

Six service groups

Buy each group on its own, or combine them into a single operations contract. Most clients start with one group and expand from there.

Database administration

Three service levels

Six database families, Oracle included. Monitoring, backup and performance tuning to an agreed commitment.

  • Oracle Database · PostgreSQL · MySQL and MariaDB · MongoDB · Redis and Dragonfly · Kafka and Redpanda
  • Monitoring with alerts, automated backup, security patching in line with advisories
  • Scheduled performance tuning and data architecture advice

Compared with hiring an in-house administrator

During leave or sick days No cover Someone is always on duty
When they resign All system knowledge is lost No impact — every system is documented
Technology coverage Usually fluent in one or two Six database families
Time to have someone in place Months of recruitment Within the week

Limits to be aware of An in-house person has three advantages no outside service can replace: they understand your organisation's business processes, they sit with the users, and they are there when called. For systems bound tightly to internal processes, the best arrangement is usually one person inside, with us underneath as the foundation.

Business email

Priced by system

Mailboxes on your own domain, built on iRedMail, with data held in Vietnam and no charge per mailbox.

  • Webmail, with synchronisation to Outlook and mobile mail apps
  • SPF, DKIM and DMARC configured — the three records that decide whether mail reaches the inbox
  • Migration from your existing system, running in parallel until you are certain

Compared with an overseas email service

How it is charged Per mailbox, rising as you add people Per system; adding mailboxes adds no licence cost
Where data is held Overseas Vietnam
If you want to stop the service Subject to the supplier's policy The system and the data belong to your organisation

Limits to be aware of Open source spam filtering is not as strong as Google's, and the overseas services bundle an office suite that iRedMail does not have. If your organisation makes heavy use of those tools, the actual saving will be smaller.

Backup & disaster recovery

Drills on a schedule

Backup alone is not enough. We run recovery drills on a schedule and hand over a written record each time.

  • Test restores into an isolated environment, with the whole process timed
  • A record of the actual recovery time and the data point recovered to
  • Backups protected against modification and deletion, and against ransomware
  1. Live copy

    On the production system

    Not a backup — this is the original data

  2. Near copy

    Same data centre

    Accidental deletion, application faults, fast recovery

  3. Far copy

    A different data centre, write-protected

    Fire, ransomware, region-wide failure

The difference lies in a step that is not on the diagram: periodically restoring from the far copy into an isolated environment, timing it, and handing over the record. A backup that has never been test-restored is not yet known to work.

Limits to be aware of If a test restore fails or runs slower than the commitment, the record says exactly that. A clean record that is wrong helps no one on the day the system actually fails.

Systems operations

By project or monthly

Servers, web servers, firewalls, network equipment and centralised monitoring.

  • Operating system installation and hardening, with scheduled patch management
  • Nginx, Apache, IIS · auto-renewing TLS certificates · load balancing
  • System security review: a report of configuration weaknesses ranked by priority

Limits to be aware of This is a configuration review, not penetration testing. An organisation that needs penetration testing for compliance should engage a firm that specialises in it.

Virtual servers, S3 storage, private cloud

Sized to actual load

Infrastructure in domestic data centres, with security configuration and operations included.

  • Virtual servers sized to actual load, not sold as fixed packages
  • Object storage compatible with the S3 protocol
  • Private cloud for organisations that require full isolation

Limits to be aware of We do not compete on the lowest rental price. At small configurations, international providers are usually cheaper. If your organisation already has its own operations team, buying directly from an infrastructure provider makes more sense.

Oracle to PostgreSQL migration

Four steps, stoppable partway

A feasibility assessment first, then a pilot migration, and only then the full move. You can stop at any step.

  • Converts almost automatically: tables, indexes, constraints, foreign keys
  • Must be rewritten: PL/SQL packages, complex triggers, Oracle-specific functions
  • Parallel running to reconcile business results before the real cutover
  1. Feasibility assessment

    We inventory the schema, count the objects that must be rewritten, and review applications that depend on Oracle syntax. The result is a report on the volume of work, with a recommendation on whether to migrate.

    You can stop here — what has been delivered remains usable, including if you appoint another contractor to continue

  2. Pilot migration

    We pick one module with clear boundaries, migrate it completely, and measure the real outcome. This step confirms whether the estimates from step one hold.

    You can stop here — what has been delivered remains usable, including if you appoint another contractor to continue

  3. Parallel running and reconciliation

    Both systems run at once on the same data. Results are reconciled at the business level, not merely by counting records. Every discrepancy must be explained before going further.

  4. Full migration and hand-over

    We migrate everything, keeping the Oracle system in read-only mode for an agreed period so there is a way back. We hand over operating documentation and train your organisation's technical team.

Limits to be aware of A system near the end of its life, running stably, or tied to applications written against Oracle syntax usually should not be migrated. The assessment report may well conclude that you should not proceed.

Commitment

Three service commitment levels

Response time runs from the moment we receive the incident report, not from the moment the incident occurs. Many contracts leave this point vague.

Level Response time Support hours
Basic 8 working hours Monday – Friday, 8:00–17:30
Standard 4 hours Office hours, with rostered out-of-hours cover
Advanced 1 hour 24/7, including weekends and public holidays

No dependence on one individual

Every system we take on has written documentation and at least two engineers who know it.

Rostered cover with hand-over

Outside office hours, engineers cover by roster, not by individual. Each shift change includes a hand-over recording system status and work in progress, so the incoming engineer does not have to start by asking.

No client lock-in

We hand over the full configuration and documentation. If you want to operate it yourself or move to another supplier, you can do so immediately.

Scope written into the contract

What is included and what is charged separately are set down in writing from the outset. Scope is not left to be argued over during an incident.

At the Standard level, a rostered engineer receives and handles incidents outside office hours. At the Advanced level, the commitment applies continuously, including weekends and public holidays. How incident severity is classified is set out in the contract so both sides agree in advance, rather than arguing during an incident.

Start with an infrastructure review

We inventory your servers, databases, backup status and the risks already present, then hand over a report ranked by priority.