| 일 | 월 | 화 | 수 | 목 | 금 | 토 |
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | ||
| 6 | 7 | 8 | 9 | 10 | 11 | 12 |
| 13 | 14 | 15 | 16 | 17 | 18 | 19 |
| 20 | 21 | 22 | 23 | 24 | 25 | 26 |
| 27 | 28 | 29 | 30 |
- 저지연시스템
- KRX HFT
- hft개발
- speed trading
- high frequency trading
- Korea krx
- futures options
- ultra low latency
- mmlp
- DMA개발
- 자동매매
- algo trading
- fep개발
- HFT system
- 주문 오더북
- 시스템 트레이딩
- hft tuning
- trading engineering
- 트레이딩 엔지니어링
- korea inbound
- low latency
- dma hft
- 주문fep
- 프랍데스크
- Arbitrage Trading
- Korea dma
- low latency trading
- system trading
- dma fep
- dma development
- Today
- Total
목록전체 글 (44)
Fontes Fintech
저지연 트레이딩 시스템에서 NIC Offload와 Kernel Networking이 중요한 이유Trading Engineering — Deep Dive #7저지연 트레이딩 시스템에서 네트워크 성능을 이야기할 때 흔히 CPU, Kernel, Socket, TCP/UDP에 집중합니다.하지만 그보다 먼저 데이터를 받아들이는 장치가 있습니다.바로 NIC(Network Interface Card) 입니다.NIC는 단순히 네트워크 케이블을 연결하는 장치가 아닙니다.최근의 고성능 NIC는Packet Receive / TransmitDMAChecksum 계산Packet SegmentationReceive Queue 분배RSSMulti-QueueInterrupt 생성등 다양한 기능을 하드웨어에서 처리할 수 있습니다.따라..
Kernel Bypass와 User-Space Networking이 저지연 트레이딩 시스템에서 중요한 이유Trading Engineering — Deep Dive #8저지연 트레이딩 시스템에서 네트워크 latency를 줄이려고 하면 자연스럽게 다음과 같은 질문에 도달하게 됩니다."애플리케이션이 패킷을 받기 위해 반드시 Kernel을 거쳐야 할까?"일반적인 Linux networking에서는 다음과 같은 경로를 사용합니다.NIC ↓Driver ↓Linux Kernel ↓Network Stack ↓Socket ↓Application이 구조는 매우 안정적이고 범용적입니다.하지만 극도로 낮은 latency가 필요한 시스템에서는 Kernel을 거치는 과정 자체가 overhead가 될 수 있습니다.이러한 문제를 해..
Interrupts, Polling and Network Latency in Low-Latency Trading SystemsTrading Engineering — Deep Dive #6In a low-latency trading system, receiving a packet is only the beginning.A market data packet or an order response must travel through multiple layers before the application can actually process it.A simplified path looks like this:NIC ↓Interrupt / Polling ↓Kernel ↓Network Stack ↓Socket ↓Appl..
저지연 트레이딩 시스템에서 Interrupt, Polling과 Network Latency가 중요한 이유Trading Engineering — Deep Dive #6저지연 트레이딩 시스템에서 네트워크 패킷을 받는 것은 시작에 불과합니다.시장 데이터 패킷이나 주문 응답은 실제 애플리케이션이 처리하기까지 여러 계층을 거쳐야 합니다.단순화하면 다음과 같습니다.NIC ↓Interrupt / Polling ↓Kernel ↓Network Stack ↓Socket ↓Application 각 단계에서 latency가 발생할 수 있습니다.그리고 들어오는 패킷을 어떻게 감지하고 처리하는가에 따라 latency와 CPU 사용량이 크게 달라질 수 있습니다.따라서 저지연 트레이딩 시스템을 설계할 때는 Interrupt, Pol..
Why NUMA and Memory Locality Matter in Low-Latency Trading SystemsTrading Engineering — Deep Dive #4In low-latency trading systems, CPU performance is only part of the story.Where the data is located can be just as important.Modern servers may use multiple CPU sockets and NUMA (Non-Uniform Memory Access) architectures. In such systems, memory access does not always have the same cost.Even within..
epoll vs. Busy Polling in Low-Latency Trading SystemsTrading Engineering — Deep Dive #5When developing network applications on Linux, epoll() is one of the most commonly used mechanisms for handling I/O events.It is highly efficient for general-purpose servers.However, low-latency trading systems sometimes require a different approach.When market data or order messages must be processed with ext..
