From infrastructure fundamentals to Canadian regulatory considerations, explore the topics that matter when evaluating hosting environments.
The Uptime Institute tier classification system provides a standardized way to describe data centre reliability and redundancy. Understanding what each tier actually means helps you evaluate whether a facility's capabilities match your requirements.
Single path for power and cooling distribution. No redundant components. Vulnerable to disruptions from planned and unplanned events. Annual maintenance requires full shutdown. Appropriate for non-critical workloads where downtime is acceptable.
Includes redundant capacity components (N+1) but still single distribution path. Partial redundancy reduces some risks. Maintenance still affects operations. More resilient than Tier I but planned downtime remains necessary.
Multiple distribution paths with one active. All equipment dual-powered. Planned maintenance can occur without shutdown. Unplanned events may still cause disruption. Common choice for business-critical applications requiring high availability.
Multiple active distribution paths. Compartmentalized to prevent single events from affecting entire facility. Can sustain equipment failure without impact to operations. Highest reliability level with corresponding cost implications.
Reliable power delivery involves utility connections, backup generators, uninterruptible power supplies, and distribution systems. The configuration of these components determines how equipment responds to power disruptions.
Primary power source for data centres. Facilities may have single or multiple utility feeds from different substations. Diverse utility feeds reduce vulnerability to upstream utility failures but involve additional infrastructure cost.
Uninterruptible power supplies provide immediate backup when utility power fails, bridging the gap until generators start. Battery capacity, configuration (centralized vs distributed), and maintenance requirements all affect reliability.
Diesel or natural gas generators provide extended backup power. Generator capacity, fuel storage, automatic transfer switches, and testing procedures determine how long a facility can operate during extended utility outages.
N+1 means one extra component beyond minimum. 2N means fully duplicated systems. Redundancy level affects both reliability and cost. Understanding your actual uptime requirements helps determine appropriate redundancy.
Servers generate heat that must be continuously removed. Cooling approaches vary in efficiency, capacity, and redundancy. Modern high-density computing has driven evolution in cooling technology.
Computer Room Air Conditioning units are traditional approach. Refrigerant-based cooling with raised floor or overhead distribution. Established technology with known maintenance requirements but less efficient than newer approaches.
Cooling units placed directly in server rows bring cooling closer to heat source. More efficient than perimeter cooling for high-density environments. Allows for targeted cooling of hot spots.
Hot aisle or cold aisle containment separates hot and cold air streams. Improves cooling efficiency by preventing mixing. Can significantly reduce cooling costs in properly designed implementations.
Uses outside air when ambient temperature allows. Reduces mechanical cooling load and energy consumption. Particularly effective in cooler Canadian climates. Requires appropriate filtration and humidity control.
Network connectivity affects application performance, redundancy options, and costs. Understanding carrier options, bandwidth pricing, and latency considerations helps in facility selection.
Carrier-neutral facilities allow choice of multiple network providers. Enables competitive pricing and diverse connectivity. Some facilities have exclusive carrier arrangements that limit options.
Direct connection to internet exchanges improves routing efficiency and reduces latency. Toronto, Montreal, and Vancouver host major Canadian exchanges. Proximity to exchange points can affect application performance.
Many Canadian businesses need connectivity to US markets. Understanding cross-border routing, latency to major US cities, and data transit regulations affects facility choice for certain use cases.
Bandwidth pricing varies by market and provider. Blended transit combines multiple carriers for redundancy and cost optimization. Understanding pricing models helps budget for connectivity costs.
Canadian privacy laws and data sovereignty considerations affect hosting decisions for many organizations. Understanding regulatory requirements helps determine whether Canadian-based hosting is necessary.
Personal Information Protection and Electronic Documents Act governs private sector data handling across Canada. Organizations must understand how hosting location affects compliance obligations and cross-border data transfer restrictions.
Quebec, British Columbia, and Alberta have provincial privacy legislation. Requirements vary by jurisdiction. Organizations operating across provinces need to understand overlapping and differing requirements.
Some regulations or contracts require data to remain within Canadian borders. Healthcare, financial services, and government sectors often have specific data residency requirements that affect facility selection.
Data stored outside Canada may be subject to foreign legal requests. Canadian hosting can provide additional protection from certain foreign jurisdiction claims, though legal complexity exists in this area.
Dive deeper into tier classification systems and how they apply to facility evaluation.