Challenges of Physical Clocks in Distributed Systems

Challenges of Physical Clocks

• Physical clocks naturally drift over time, even with synchronization efforts.
• Geographical dispersion affects clocks in different locations, experiencing different time due to relativity.
• Limited causality means physical timestamps don't guarantee the order of events in a distributed system.

Logical Clocks

• Establish causal relationships between events, independent of physical time.
• Focus on ordering events based on causality, not absolute time.
• Examples include Lamport Clocks (simple and efficient, establish partial order) and Vector Clocks (more complex, handle concurrent events and provide stronger ordering guarantees).

UTC Time and Leap Seconds

• UTC combines the accuracy of International Atomic Time (TAI) with the Earth's rotation.
• Leap seconds are added to UTC to maintain consistency with astronomical time (Earth's rotation).
• Challenges of leap seconds include disrupting computer systems, especially Unix time (epoch time), and introducing complexity.

Implications for Distributed Systems

• Importance of consistent time: processors in a distributed system need a uniform notion of time to maintain consistent state.
• Skew and drift can significantly impact time-dependent operations in distributed systems.

Atomic Clocks

• Highly accurate clocks based on atomic properties.
• Defined the second based on cesium atom transitions, more precise than quartz clocks (drift of 1 second in 300 million years).

Need for Precise Clocks

• High-frequency trading requires microsecond accuracy for timestamps.
• Regulations like MiFID II set specific timing requirements for financial systems.
• Applications in defense, geology, astronomy, and navigation also rely on precise time.

Time Representation in Computers

• Epoch time (Unix time): seconds elapsed since January 1, 1970.
• ISO 8601: standard format for date and time.

Synchronization Algorithms for Physical Clocks

• External Synchronization: relies on an external, authoritative time source to synchronize the clocks of various nodes in a distributed system.
• Internal Synchronization: nodes in the system synchronize their clocks directly with each other, without relying on an external source.
• NTP (Network Time Protocol):
  • Uses a hierarchy of servers, with primary servers receiving time directly from atomic clocks and secondary servers synchronizing with primaries.
  • Strengths of NTP: reliable, secure, and utilizes a stratum hierarchy with redundant servers and paths.

Distributed Systems

• Characteristics:
  • No global physical clocks.
  • Autonomous processors/independent processors/independently failing.
  • No global shared common memory.
• Channels and Links: communication pathways used for message transmission between processes.

Explore the challenges faced by physical clocks in distributed systems, including drift, geographical dispersion, and limited causality. Learn how logical clocks provide a solution by establishing causal relationships between events.