A 100,000‑mailbox, three-region Exchange platform with automatic database failover
An email provider that hired our team to build its in-house Microsoft Exchange infrastructure.
- Scale
- 100,000+ mailboxes, 500+ domains, 250 TB
- Engagement
- Design, build and operation
The challenge
The provider needed carrier-grade email of its own, with no single point of failure, every patch applied fast and deliverability customers could trust. Mail was its product, so an outage was a product outage for every customer at once.
At this scale, small failures become daily events, so the design assumed they would happen. Database copies on standby servers take over automatically when a primary fails, and three regions mean that no single facility is a single point of failure.
Operations were as important as architecture. Applying every Microsoft patch within four hours of release, authenticating mail with SPF, DKIM and DMARC, and filtering spam in both directions kept the platform both secure and trusted by the receiving servers that decide whether customers' mail is delivered.
What our team delivered
- Redundant dedicated colocation in three regions, 250 TB of iSCSI storage and virtualized servers
- Exchange database availability groups with copies in more than one region, so a server or a site can fail without users losing mail access
- Automatic database failover to standby copies for near-zero downtime for every user
- Postfix, Active Directory, Microsoft 365 and Exchange 2019 through Exchange Server Subscription Edition
- Every Microsoft patch applied within 4 hours of release, through a staged process that kept copies in service during updates
- SPF, DKIM, DMARC and TLS encryption on inbound and outbound mail; custom inbound and outbound spam filters
- Custom monitoring agents with an in-house algorithm that watches message patterns
- Autodiscover for 500+ domains; certbot integration and scripts that renew SSL certificates automatically
- Custom APIs integrating e-commerce and the messaging infrastructure
Result
Carrier-grade email for 500+ domains with no single point of failure. The same architecture, at a smaller scale, now protects the email of the conglomerate, PR agency and consultancy described on this page, and it is the basis of our Exchange and Active Directory practice.
Mail that keeps flowing
Incoming mail is authenticated and scanned before it reaches anyone, mailboxes are copied across three regions so a server failure goes unnoticed, and outgoing mail leaves signed and encrypted.
DMARC: from watching to rejecting
Share of mail claiming the domain that passes DMARC (SPF or DKIM aligned), week by week. In the monitoring phase (p=none) reports reveal forgotten senders; each one gets SPF and DKIM fixed. Quarantine sends failures to spam, reject refuses them at the receiving server. Illustrative figures.
Twelve months of availability
Measured availability per site over twelve months against a 99.99% target, which allows 52.6 minutes of downtime a year. Six of seven sites meet it; Singapore misses and is shown in red. Illustrative figures.
Facing something similar?
Tell us about your environment. We will tell you plainly what we would do and what it would take.