Data Center Controls

Prior to choosing a location, AWS performs initial environmental and geographic assessments. Data center locations are carefully selected to mitigate environmental risks, such as flooding, extreme weather, and seismic activity. Our Availability Zones are built to be independent and physically separated from one another.

Data centers are designed to anticipate and tolerate failure while maintaining service levels. In case of failure, automated processes move traffic away from the affected area. Core applications are deployed to an N+1 standard, so that in the event of a data center failure, there is sufficient capacity to enable traffic to be load-balanced to the remaining sites.

AWS has identified critical system components required to maintain the availability of our system and recover service in the event of outage. Critical system components are backed up across multiple, isolated locations known as Availability Zones. Each Availability Zone is engineered to operate independently with high reliability. Availability Zones are connected to enable you to easily architect applications that automatically fail-over between Availability Zones without interruption. Highly resilient systems, and therefore service availability, is a function of the system design. Through the use of Availability Zones and data replication, AWS customers can achieve extremely short recovery time and recovery point objectives, as well as the highest levels of service availability.

AWS continuously monitors service usage to deploy infrastructure to support our availability commitments and requirements. AWS maintains a capacity planning model that assesses our infrastructure usage and demands at least monthly. This model supports planning of future demands and includes considerations such as information processing, telecommunications, and audit log storage.

The AWS Business Continuity Plan outlines measures to avoid and lessen environmental disruptions. It includes operational details about steps to take before, during, and after an event. The Business Continuity Plan is supported by testing that includes simulations of different scenarios. During and after testing, AWS documents people and process performance, corrective actions, and lessons learned with the aim of continuous improvement.

AWS incorporates pandemic response policies and procedures into its disaster recovery planning to prepare to respond rapidly to infectious disease outbreak threats. Mitigation strategies include alternative staffing models to transfer critical processes to out-of-region resources, and activation of a crisis management plan to support critical business operations. Pandemic plans reference international health agencies and regulations, including points of contact for international agencies.

AWS provides physical data center access only to approved employees. All employees who need data center access must first apply for access and provide a valid business justification. These requests are granted based on the principle of least privilege, where requests must specify to which layer of the data center the individual needs access, and are time-bound. Requests are reviewed and approved by authorized personnel, and access is revoked after the requested time expires. Once granted admittance, individuals are restricted to areas specified in their permissions.

Third-party access is requested by approved AWS employees, who must apply for third-party access and provide a valid business justification. These requests are granted based on the principle of least privilege, where requests must specify to which layer of the data center the individual needs access, and are time-bound. These requests are approved by authorized personnel, and access is revoked after request time expires. Once granted admittance, individuals are restricted to areas specified in their permissions. Anyone granted visitor badge access must present identification when arriving on site and are signed in and escorted by authorized staff.

Physical access to data centers in AWS GovCloud (US) is restricted to employees who have been validated as being US citizens.

Access to data centers is regularly reviewed. Access is automatically revoked when an employee’s record is terminated in Amazon’s HR system. In addition, when an employee or contractor’s access expires in accordance with the approved request duration, his or her access is revoked, even if he or she continues to be an employee of Amazon.

Physical access to AWS data centers is logged, monitored, and retained. AWS correlates information gained from logical and physical monitoring systems to enhance security on an as-needed basis.

We monitor our data centers using our global Security Operations Centers, which are responsible for monitoring, triaging, and executing security programs. They provide 24/7 global support by managing and monitoring data center access activities, equipping local teams and other support teams to respond to security incidents by triaging, consulting, analyzing, and dispatching responses.

Physical access points to server rooms are recorded by Closed Circuit Television Camera (CCTV). Images are retained according to legal and compliance requirements.

Physical access is controlled at building ingress points by professional security staff utilizing surveillance, detection systems, and other electronic means. Authorized staff utilize multi-factor authentication mechanisms to access data centers. Entrances to server rooms are secured with devices that sound alarms to initiate an incident response if the door is forced or held open.

Electronic intrusion detection systems are installed within the data layer to monitor, detect, and automatically alert appropriate personnel of security incidents. Ingress and egress points to server rooms are secured with devices that require each individual to provide multi-factor authentication before granting entry and badge before exiting. These devices will sound alarms if the door is forced open without authentication, held open, or opened for exit during an emergency. Door alarming devices are also configured to detect instances where an individual enters a data layer without providing multi-factor authentication or exits without proper badging. Alarms are immediately dispatched to the 24/7 AWS Security Operations Centers for immediate logging, analysis, and response.

AWS assets are centrally managed through an inventory management system that stores and tracks owner, location, status, maintenance, and descriptive information for AWS-owned assets. Following procurement, assets are scanned and tracked, and assets undergoing maintenance are checked and monitored for ownership, status, and resolution.

Media storage devices used to store customer data are classified by AWS as Critical and treated accordingly, as high impact, throughout their life-cycles. AWS has exacting standards on how to install, service, and eventually destroy the devices when they are no longer useful. When a storage device has reached the end of its useful life, AWS decommissions media using techniques detailed in NIST 800-88. Media that stored customer data is not removed from AWS control until it has been securely decommissioned.

Our data center electrical power systems are designed to be fully redundant and maintainable without impact to operations, 24 hours a day. AWS ensures data centers are equipped with back-up power supply to ensure power is available to maintain operations in the event of an electrical failure for critical and essential loads in the facility.

AWS data centers use mechanisms to control climate and maintain an appropriate operating temperature for servers and other hardware to prevent overheating and reduce the possibility of service outages. Personnel and systems monitor and control temperature and humidity at appropriate levels.

AWS data centers are equipped with automatic fire detection and suppression equipment. Fire detection systems utilize smoke detection sensors within networking, mechanical, and infrastructure spaces. These areas are also protected by suppression systems.

In order to detect the presence of water leaks, AWS equips data centers with functionality to detect the presence of water. If water is detected, mechanisms are in place to remove water in order to prevent any additional water damage.

AWS monitors and performs preventative maintenance of electrical and mechanical equipment to maintain the continued operability of systems within AWS data centers. Equipment maintenance procedures are carried out by qualified persons and completed according to a documented maintenance schedule.

AWS monitors electrical and mechanical systems and equipment to enable immediate identification of issues. This is carried out by utilizing continuous audit tools and information provided through our Building Management and Electrical Monitoring Systems. Preventative maintenance is performed to maintain the continued operability of equipment.

The AWS Security Operations Center performs regular threat and vulnerability reviews of data centers. Ongoing assessment and mitigation of potential vulnerabilities is performed through data center risk assessment activities. This assessment is performed in addition to the enterprise-level risk assessment process used to identify and manage risks presented to the business as a whole. This process also takes regional regulatory and environmental risks into consideration.

Third-party testing of AWS data centers, as documented in our third-party reports, ensures AWS has appropriately implemented security measures aligned to established rules needed to obtain security certifications. Depending on the compliance program and its requirements, external auditors may perform testing of media disposal, review security camera footage, observe entrances and hallways throughout a data center, test electronic access control devices, and examine data center equipment.