1
point
fh973
Joined 1,662 karma
Co-founder / CTO Quobyte (https://www.quobyte.com)
- fh973Checked how to receive those with SDR. Turns out they are very low power and you need to basically touch the tire. Also the transmit in minute intervals. Bit exactly a a smoking gun in terms of mass surveillance.
- QUOBYTE | Santa Clara, CA and Berlin, Germany | Full-time / Remote | ONSITE | https://www.quobyte.com/
At Quobyte we are working on a highly scalable and fault-tolerant software storage system built around a parallel file system core. Our customers use us for large scale AI and HPC clusters in the enterprise and research, k8s and OpenStack infrastructures, and as a scalable backend for SaaS products. There are Quobyte clusters which span tens of thousands of machines and slurp 100s of GB/s!
Under the hood, we have built a full-stack fault-tolerant parallel file system, with everything from kernel development over our own replicated database system design to distributed algorithms (Paxos!) and performance. In short: lots of real-world challenging and fun problems!
We work as a highly efficient engineering team, ship frequently, do code reviews, and have lots of unit and integration testing. If you’re passionate about systems, we might be the right place for you!
Berlin, Germany:
* Software Engineer (with a passion for Systems)
* DevOps Engineer (for on premises engineering and support infrastructure)
US
* Sales Engineer East Coast, West Coast (remote)
For detailed job descriptions please and application process, please visit https://www.quobyte.com/company/careers or write to work at quobyte.com.
- Queues tend to be always full or always empty (see queueing theory). There is no steady state with a half full queue.
For NVMe in particular you will have a hard time filling their queues. Your perceived performance is mostly latency, as there is hardly an application that can submit enough concurrent requests.
- Aren't panels commodotized? Many western countries could probably spin up production at scale if needed and economically viable. But why do so now?
Also why talk only about panels. Inverters are cost wise in the same order of magnitude, but much more complex artifacts.
- QUOBYTE | Santa Clara, CA and Berlin, Germany | Full-time | ONSITE | https://www.quobyte.com/ At Quobyte, we are working on the Data Center File System (DCFS), a highly scalable and fault-tolerant software storage system built around a parallel file system core. Customers use Quobyte DCFS for ML and HPC clusters in industry and academia, k8s and OpenStack infrastructures, streaming, web mail, post production, origin and CGI clusters, and as a scalable backend for SaaS products. There are Quobyte DCFS clusters which span thousands of machines and crunch 100s of GB/s!
Under the hood, DCFS is a full-stack storage system, written in Java and C++. The engineering work covers everything from kernel development over our own replicated database system design to distributed algorithms (Paxos!) and performance. In short: lots of real-world challenging and fun problems! We work as a highly efficient engineering team, and do mandatory code reviews, lots of unit and integration testing,
Berlin, Germany:
* Software Engineer
* Infrastructure / Devops Engineer
East and West Coast:
* Sales Engineer
For detailed job descriptions please and application process, please visit https://www.quobyte.com/company/careers or write to work at quobyte.com.
- Donald, is it you?
- Mobileeye in Munich https://www.mobileye.com/blog/self-driving-robotaxi-sixt-ger...
Moia (Volkswagen) in Hamburg https://www.moia.io/en
Mercedes autonomous driving https://group.mercedes-benz.com/innovations/product-innovati...
- Docker swarm is great on single servers. Apparently still no such thing for Podman.
Even if the tech is not top notch, Docker got a few things right on product management.
- Max. 16Mbit in Berlin-Schöneberg here.
- Indeed, with the tmpfs move (tmp in RAM) however it sounds like they have more Desktops in mind.
You don't want to use RAM for tmp files for which you probably can't do capacity planning, and you don't to enable swap on server either.
- Swap on servers somewhat defeats the purpose of ECC memory: your program state is now subject to complex IO path that is not end-to-end checksum protected. Also you get unpredictable performance.
So typically: swap off on servers. Do they have a server story?
- Marketing material says NVIDIA DGX™ OS, which at version 7 would be an Ubuntu 24.04: https://docs.nvidia.com/dgx/dgx-os-7-user-guide/introduction...
- Ordered one in spring. Delivery time was pushed from July to September. Apparently they had a bug in the HDMI output.
- Zeppelin is operating daily sightseeing flights in several areas of Germany https://zeppelinflug.de/en/zeppelin-flights
- At Quobyte (https://www.quobyte.com) we use FUSE for the client for parallel file system access.
You can get dozens of GB/s out of FUSE nowadays. This will even improve in the near future as FUSE is adding io_uring support for communication with the kernel (instead of a pipe).
- 1 soldier in the battlefield needs several soldiers as support (logistics etc.).
- Them? Nvidia and AMD? RTX 7xxx would then be based on an old Intel process? Would buy these?