Vinay Banakar

@VinaySBanakar

HPE RnD Labs, interested broadly in systems. HPC | Privacy | Storage | Datacenters. Remote Research fellow at UT Austin | PESIT CSE.

Bangalore
Vrijeme pridruživanja: siječanj 2014.
Rođen/a 12. listopada 1995.

Tweetovi

Blokirali ste korisnika/cu @VinaySBanakar

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @VinaySBanakar

  1. Prikvačeni tweet
    4. lis 2019.

    Excited to announce GDPRBench!! A new benchmark that introduces GDPR workloads in order to quantify GDPR compliance and its impact on storage systems. (1/N)

    Prikaži ovu nit
    Poništi
  2. 18. sij

    I don't encourage sleeping 9 hours everyday but when you do after months it's the best!! Instantly makes you feel fit and healthy.

    Poništi
  3. 11. sij

    Samsung SSDs have big SLC NAND flash to meet advertised numbers however they use much slower TLC and QLC flash for actual storage. Once SLC flash is full effective write performance drops to that of TLC and QLC. Also they don't have supercap.

    Prikaži ovu nit
    Poništi
  4. 11. sij

    TIL: Enterprise SSDs have supercaps with sufficient power that they will flush the DRAM cache to flash on power failure, ensuring that nothing committed would be lost unlike consumer SSDs which also have extensive write-caching and lie about fsync but has poor powerfail safe.

    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet
    6. sij

    New year, new rights. Here's everything you need to know about the California Consumer Privacy Act.

    Poništi
  6. 28. pro 2019.

    A 2Tb persistent array will require around 32Gb or page structures! :o

    Prikaži ovu nit
    Poništi
  7. 28. pro 2019.

    With terabytes of persistent memory becoming the norm here is an interesting deliberation on where or how struct pages should be stored and accessed or if they are needed at all in Linux. Today for every 4 KB page there is a 64 B struct.

    Prikaži ovu nit
    Poništi
  8. 12. pro 2019.

    Almost two decades in and this critique still seems 'relevant'. "Systems Software Research is Irrelevant"

    Poništi
  9. proslijedio/la je Tweet
    7. pro 2019.

    We need some help: we received a record number of 43 artifacts for the initiative at & need more evaluators! If you are interested to help or would like to nominate your colleagues, please contact us before Dec 11! More info:

    Poništi
  10. proslijedio/la je Tweet
    6. pro 2019.

    I am on the tenure-track job market! I work at the intersection of systems and disciplines such as economics (PhD) and law (post-doc). I would love to join a dept that values interdisciplinary work! Application material:

    Prikaži ovu nit
    Poništi
  11. proslijedio/la je Tweet

    BIG NEWS EVERYONE! , , and I have been working really hard on something awesome! Subscribe and listen to our teaser podcast episode... and there will be more details to come tomorrow, I can't wait!! STAY TUNED!!

    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    6. stu 2019.

    This week’s cover shows Nature’s publication record over 150 years. Explore the growing web of collaboration and science in an interactive graphic here:

    Poništi
  13. proslijedio/la je Tweet
    1. stu 2019.

    From Prof. Mark Silberstein, we have a recap of USENIX ATC 2019 from an architecture perspective.

    Poništi
  14. proslijedio/la je Tweet
    28. lis 2019.

    Fabulous Artifact Evaluation program! 23 of 38 papers became AE submissions. Distinguished Artifact Awardees announced including one of the best papers! Thanks to @SupreethShastri for organizing such a great result!

    Poništi
  15. 8. lis 2019.

    I usually am a staunch supporter of experiments, but here is a nice article by Prof Jeffrey Ullman which emphasizes that experiments shouldn't be a substitute for deep understanding and analysis and lack of experiments shouldn't warrant a paper rejection

    Poništi
  16. 4. lis 2019.

    We also found that RDBMS like Postgres, not only performs better but also is relatively easy to make it GDPR compliant (naïve) than NoSQL systems like Redis. To learn more insights like these, checkout our preprint on Arxiv! (5/5)

    Prikaži ovu nit
    Poništi
  17. 4. lis 2019.

    GDPRBench provides a quantifiable measure of correctness and performance of datastores that hold personal data. It intends to help system designers realize the GDPR implications on their design choices and allows them to explore performance-compliance tradeoffs in their setup.

    Prikaži ovu nit
    Poništi
  18. 4. lis 2019.

    However, the policy has also introduced undesired effects in storage systems such as metadata explosion and serious performance degradation. We observed up to 5x & 2x drop in throughput in compliant Redis and Postgres accordingly and about 6x storage space overhead with indexing.

    Prikaži ovu nit
    Poništi
  19. 4. lis 2019.

    Although by law GDPR is limited to Europe, it has become the bedrock for new privacy legislations worldwide. It has instigated a major trend where companies have repositioned their products to be more privacy-focused and has escalated privacy awareness globally.

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    3. lis 2019.

    Happy to share our work on GDPRBench, a new benchmark similar to YCSB that helps developers understand how GDPR affects database systems. This is joint work with @SupreethShastri, , , and the school!

    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·