resource & container

來自專欄 Eva-wang

Senior Staff Engineer/Architect of Sigma -USA, Hangzhou, Beijing

Job description:

To be more specific (you are more than welcome if you are interested in one or more challenges described below):

• 1. Enable Sigma to response more quickly to resource requests from more than dozens of business units and even more business scenarios with proper resource allocation and constraints compliance. Design a more generic architecture to enable Sigma coping with the ever-growing scale in terms of both business scenario and machine number.
• 2. Manage heterogeneous resources with Sigma, such as GPU, FPGA, RDMA, QAT, etc. and support innovating techniques such as computing-storage separation, Function-as-a-Service, edge-computation, stream-computation, etc.
• 3. Improve the stability, reliability and robustness of Sigma so that the user facing services are not affected (High Availability) by better handling accidents of various causes, including: hardware/software failure, miss operation from engineers and even some catastrophes.
• 4. Based on the huge amount of information accumulated in our logs and monitors, analyze the interference between container instances, and help Sigma make better scheduling and rescheduling decision based on observation and analysis of real time resource (CPU/Mem/Network IO/etc.) consumption, app response time, and resource contention, etc. to achieve higher application performance and resource utilization rate.
• 5. Based on current status and prediction, help making machine purchase decisions including hardware type/configuration, network structure, and deployment plans in datacenters, etc. Also, helping us design a heuristic charging and price policy for container services to guide our internal business unit users towards optimal resource utilization and lower cost.

.We are looking for future colleagues for tech leaders, architect, core developers, researchers, etc. The skill stack is listed as follows:

• 1. Good understanding of virtualization and/or container technology. Knowing the mechanism of namespace, cgroup, etc. in Linux. Having solid knowledge in resource scheduling/management field and being familiar with typical existing resource management frameworks, their structure, rational & concepts underlying their design. The experience of developing large scale resource management system would be a big plus.
• 2. Good understanding of IDC planning, network structure, server deployment, and good knowledge with CPU, memory, disk/storage. Familiar with typical computation patterns such as batch, real-time, streaming, etc. Knowledge and hands-on machine learning skills would be a big plus.
• 3. We would prefer candidates to have solid background in Golang and be able to handle problem independently.
• 4. Great expression & communication skills, and team work sprit. Being able to work with multiple teams. Good insights in technology and business evolvement and being able to make decisions on tech direction and make that happen.

Senior Staff Engineer /Architect of Pouch Container-USA, Hangzhou, Beijing

Pouch is featuring strong isolation, p2p image distribution, kernel compatibility, high performance as well as high resource efficiency. For this position, we expect candidates to improve the stability, security, portability for Pouch. To be specific (you are more than welcome if you are interested in one or more challenges described below):

• 1. Make Pouch to able response more quickly to resource requests from more than 10 business units and even more business scenarios with proper resource allocation and constraint compliance. Design a more generic architecture to enable Pouch coping with the ever-growing scale in both business scenario and machine number.
• 2. Manage heterogeneous resources with Pouch, such as GPU, FPGA, RDMA, QAT, etc. and support innovating techniques such as computing-storage separation, Function-as-a-Service, edge-computation, stream-computation, etc.
• 3. Abstract over several dozens of machines, hundreds of system-software (with several versions of kernels) and provide unified and stable container runtime for our users. Quick response to various failures, propose solutions and fix them into software distribution to be more generic whenever possible.
• 4. Implement finer grained & priority based resource isolation, including CPU, Memory, Disk IO, network IO, process handlers, etc. so that the latency critical containers could utilize resource upon request immediately and mitigate resource contention as possible.
• 5. Bring unified user experience between VM and Pouch. Improve design on transplanting existing applications that used to run in VM or physical servers to Pouch, and deal with the difference between OS/Kernel versions. Ensure compatibility with existing management tools.
• 6. Promote Pouch in open source community and make Pouch a popular choice of container runtime. Export Pouch to public cloud and other private clouds, serving more small medium enterprises and non-IT focused enterprises. Try to lead the container tech direction for the next 10 years and prepare for the next generation of container technologies.

We are looking for future colleagues for tech leaders, architect, core developers, researchers, etc. The skill stack is listed as follows:

• 1. We would prefer candidates to have solid background in Golang or C or java and be able to handle problem independently. Familiar with container open source community and Linux community. With strong interest in container and other virtualization technologies. Having a deep understanding in container technology.
• 2. Having a deep understanding on Linux and mechanism such as namespace, cgroup, etc. Familiar with implementation of resource isolation. Good knowledge on Linux kernel. Having the experience in developing Linux kernel is a big plus.
• 3. Experience such as architecture design, overseeing from design to implementation as leader independently and code contribution in open source community, would all be a big plus.
• 4. Good insights in technology and business evolvement and being able to make decisions on tech direction and make that happen.

CV email：eva.wang_hunter@foxmail.com

推薦閱讀：