dlock

Dlock

Dlock is a server for distributed locking for process synchronization & exclusive transactions in distributed systems.

Generally, a distributed lock can be used to coordinate access to a resource or interest in such a resource in a distributed environment. In other words, distributed locks are useful when multiple systems need to reach a consensus about a shared resource.

Lock Modes

As apposed to standard single-host locking mechanisms, distributed locks can be categorized into distinct modes with varying degress of distributed access guarantees.

• Null (NL) : indicates interest in the resource, but does not prevent other process from locking it
• Concurrent Read (CR) : Indicates a desire to read (but not update) the resources. Allows other processes to read or update the resource but prevents exclusive access to it.
• Concurrent Write (CW) : Indicates a desire to read and update the resource. It allows other processes to read or update the resource, but prevents others from having EX access to it.
• Protected Read (PR) : Traditional share lock, which indicates a desire to read the resource but prevents others from updating it. Others can however also read the resource.
• Protected Write (PW) : Traditional update lock, indicates a desire to read and update the resource and prevents others from updating it. Others with Concurrent read access can however read the resource
• Exclusive (EX) : Traditional exclusive lock, which allows read and update access to the resource, and prevents others from having access to it.

Notes

• A typical mutex corresponds one-to-one with a distributed exclusive lock (EX).
• There is no one-to-one mapping for a RW mutex in a distributed setting, although Protected Read & Write locks correspond closely to many of a RW mutexe's guarantees

Support Matrix

Dlock specific guarantees

Exclusive Locks (EX)

• Liveliness A : A lock is always eventually released when the process holding it crashes or exits unexpectedly.

• Liveliness B : A lock is always eventually released when its backend store is unavailable.

• Atomicity A : No two processes or threads can hold the same lock at the same time.

• Atomicity B : Any call to unlock will always eventually release the lock

References

• Distributed Lock Manager. (n.d.). In Wikipedia. Retrieved from https://en.wikipedia.org/wiki/Distributed_lock_manager
• Kleppmann, Martin. "Designing Data-Intensive Applications." (2019).
• Redis redlock algorithm from https://redis.io/docs/manual/patterns/distributed-locks/