kipjak 1.0.25

Concurrency via multithreading, multiprocessing and multihosting

s•kipja•c•k

A small, pelagic fish from the tuna family. Found in large schools in the tropical and sub-tropical oceans of the world.

---

The kipjak library is applicable to;

• Large, multiprocess website backends
• Applications with complex multithreading requirements
• Components requiring one or more subprocesses
• Complex, distributed process control systems

---

The technical goal of the kipjak library is to deliver software communications, wherever the relevant software entities may be located. At the lowest level this involves a message-driven framework providing for seamless communication between threads, processes and hosts. At higher levels it delivers responsiveness, concurrency and clarity.

• Sophistication with clarity ••• separation from the details of threads and networks
• Quick on-boarding ••• cookbook of multithreading, multiprocessing and multihosting solutions
• Optimal throughput ••• load distribution across threads, processes and hosts
• Quality client feedback ••• negative responses to requests during periods of heavy load
• Simple leveraging of sub-processes ••• load child processes as callable libraries
• Process orchestration ••• definition and execution of persisted process groups
• System daemons ••• foreground or background execution
• True application messaging ••• send and receive application values; Person, list[Person], dict[UUID,Person], bool, etc
• Broker-less communicatins ••• no middleware, no additional hop, no additional latency
• Publish-subscribe networking ••• networking without network addresses
• Industrial-grade encryption ••• NaCl (Salt)
• Automated transport monitoring ••• discreet keep-alive protocol
• HTTP integration ••• RPC or RESTful APIs
• Technical support ••• baked-in logging, log storage and log recovery

A few comparisons of related technologies with kipjak;

• HTTP is a blocking, request-response model ••• fundamentally asynchronous with HTTP integration
• ZeroMQ is broker-based middleware ••• uses direct peer-to-peer connections
• NATS delegates message serialization, e.g. to protobuf ••• sends and receives final application values