DCAT flashcards
DCAT · Build recall with a focused review session.
32 cards. Read the term and the definition here. The deck at the top of this page is optional.
- Tier IV availability
- A data center classification providing 99.995% availability with fault-tolerant, concurrently maintainable infrastructure, allowing any component to be serviced without affecting IT operations.
- PUE formula
- Power Usage Effectiveness equals total facility power divided by IT equipment power. A value of 1.0 is theoretically perfect; higher values indicate more overhead energy for cooling and losses.
- Hot aisle containment
- A physical enclosure that captures server exhaust air and returns it to the cooling units, preventing hot air from mixing with cold intake air and improving cooling efficiency.
- N+1 redundancy
- A design where one additional component beyond the required minimum is present, so a single failure does not reduce capacity below the load requirement.
- Rack unit height
- One U equals 1.75 inches or 44.45 millimeters. A standard 42U rack provides 42 units of vertical mounting space for servers and network equipment.
- ECC memory
- Error-correcting code memory detects and corrects single-bit errors and detects multi-bit errors, reducing silent data corruption on servers handling critical workloads.
- NUMA architecture
- Non-uniform memory access assigns memory locally to each processor. Local access is faster than remote access to another processor's memory, affecting workload placement and latency.
- BMC out-of-band management
- A baseboard management controller provides independent server management for power control, sensors, and console access even when the operating system is unresponsive or powered off.
- RAID 5 parity
- RAID 5 stripes data and distributed parity across at least three drives, surviving one drive failure while usable capacity equals total drives minus one times drive size.
- SAN versus NAS
- A SAN presents block-level storage over Fibre Channel or iSCSI to hosts, while NAS presents file-level shares over NFS or SMB to clients on the network.
- Thin provisioning
- Storage allocation method that presents logical capacity on demand rather than reserving physical space upfront, risking capacity exhaustion if over-subscription exceeds available physical storage.
- Synchronous replication
- A replication mode where writes are acknowledged only after the remote site confirms receipt, achieving near-zero recovery point objective at the cost of added latency.
- VLAN purpose
- A virtual LAN logically segments a Layer 2 broadcast domain on shared switching hardware, improving security and reducing broadcast traffic without requiring separate physical switches.
- Spine-leaf topology
- A data center network design where every leaf switch connects to every spine switch, providing predictable low-latency east-west traffic and easy horizontal scaling.
- Single-mode fiber
- Optical fiber with a small core that carries a single light path, supporting very long distances and high bandwidth, typically used for campus and inter-building links.
- Jumbo frames
- Ethernet frames with an MTU larger than the standard 1500 bytes, commonly 9000 bytes, reducing per-packet overhead but requiring consistent configuration across all devices on the path.
- UPS double conversion
- An online UPS continuously converts AC to DC and back to AC, isolating equipment from utility fluctuations and outages at the cost of lower efficiency.
- Watts law
- Electrical power in watts equals voltage multiplied by current in amperes. For alternating current, the result is multiplied by the power factor to get real power.
- ASHRAE inlet range
- Recommended server intake air temperature is approximately 18 to 27 degrees Celsius, balancing equipment reliability against cooling energy consumption in the white space.
- Cooling ton
- A unit of cooling capacity equal to 12,000 BTU per hour, approximately 3.517 kilowatts, historically based on the heat absorbed by melting one ton of ice daily.
- Cat6a cabling
- Twisted-pair copper cable rated for 10 Gigabit Ethernet up to 100 meters, offering improved shielding and reduced crosstalk compared with Cat6.
- ESD wrist strap
- A grounding device worn on the wrist and bonded to chassis ground that safely discharges static electricity before handling sensitive electronic components such as memory or processors.
- Hot-swap replacement
- Replacing a failed redundant component such as a drive or power supply while the system remains operational, requiring verified redundancy before removal to avoid an outage.
- Cable labeling
- Marking both ends of every cable with consistent identifiers and recording port assignments so technicians can trace connections accurately during changes or troubleshooting.
- Layer 1 troubleshooting
- Verifying physical elements first, including power, cable seating, link indicator lights, and transceiver installation, before investigating higher-layer protocol or configuration issues.
- BMC system event log
- A persistent hardware log maintained by the baseboard management controller that records power events, sensor thresholds, and failures useful for diagnosing issues when the OS is unavailable.
- SMART status
- Self-monitoring analysis and reporting technology built into drives reports health indicators such as reallocated sectors and pending failures, enabling proactive replacement before data loss.
- Baseline metrics
- Normal performance measurements captured during healthy operation, used as a reference to detect anomalies in CPU, memory, disk latency, and network utilization during troubleshooting.
- Lockout tagout
- A safety procedure that physically locks energy isolation devices and attaches warning tags, preventing accidental re-energization while maintenance is performed on electrical equipment.
- Arc flash PPE
- Personal protective equipment rated for incident energy exposure, required by NFPA 70E when working near energized electrical equipment to prevent severe burns.
- Clean agent suppression
- Fire suppression using gases such as FM-200 or Novec that extinguish fires without water residue or conductive damage, suitable for rooms containing energized IT equipment.
- Least privilege
- A security principle granting users and systems only the access required for their role, reducing the impact of compromised credentials or insider misuse in the data center.