Offloading Data Center Tax
Akshay Revankar 1 , Charan Renganathan 1 , Sartaj Wariah 1 · 2025
Read Offloading Data Center Tax by Akshay Revankar 1 , Charan Renganathan 1 , Sartaj Wariah 1 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-06558.txt.
Description
Akshay Revankar, Charan Renganathan, and Sartaj Wariah hunt for shared overhead work inside data centers that could move onto accelerator hardware. Modern data centers run varied workloads that repeat the same low-level functions, nicknamed tax components. Speeding up a tax pays off across the whole fleet, but offloading everything to accelerators is impractical.
The team profiles MongoDB, a database common in data center applications, running inside a standard benchmark suite. They identify its tax components and trace what those components demand from the processor. From those observations they suggest which groups of tax components could be offloaded together for the broadest performance gain.
Catalog record
| CREATOR | Akshay Revankar 1 , Charan Renganathan 1 , Sartaj Wariah 1 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 24 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-06558