| Paper Abstract and Keywords |
| Presentation |
2026-11-05 10:00
Wide-Area Load Balancing for LLM Inference in AI-and-RAN Environments Using Inter-Region KV Cache Migration and Pipeline Parallelism Taichi Kondo, Shota Ono, Akihiro Nakao (The Univ. of Tokyo) NS2026-121 |
| Abstract |
(in Japanese) |
(See Japanese page) |
| (in English) |
In AI-and-RAN, GPUs deployed at fixed base station sites are shared between RAN processing and AI workloads. However, AI request arrivals vary temporally and spatially, potentially causing sudden load increases in specific regions. Under high load, insufficient VRAM capacity to accommodate the Key-Value Cache (KV Cache) required for Large Language Model (LLM) inference can delay the start of inference for new requests, degrading user experience. This study proposes a wide-area load balancing method for LLM inference that combines inter-region KV Cache migration with intra-region Pipeline Parallelism (PP) to expand KV Cache capacity. Results show that KV Cache migration, the addition of two-stage PP, and their combination in the proposed method reduce average Time To First Token (TTFT) by 30.7%, 51.4%, and 78.2%, respectively, compared with a KV recomputation-based redirection method. Furthermore, the proposed method achieves the specified Service Level Objective (SLO), with TTFT remaining within 2 s for all 3,000 requests across five random seeds. |
| Keyword |
(in Japanese) |
(See Japanese page) |
| (in English) |
LLM Inference / KV Cache / Pipeline Parallelism / AI-and-RAN / / / / |
| Reference Info. |
IEICE Tech. Rep., vol. 126, no. 222, NS2026-121, pp. 7-12, Nov. 2026. |
| Paper # |
NS2026-121 |
| Date of Issue |
2026-10-29 (NS) |
| ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
| Download PDF |
NS2026-121 |
| Conference Information |
| Committee |
NS ICM NV HCL |
| Conference Date |
2026-11-05 - 2026-11-06 |
| Place (in Japanese) |
(See Japanese page) |
| Place (in English) |
Beppu B-con Plaza + Online |
| Topics (in Japanese) |
(See Japanese page) |
| Topics (in English) |
Network quality, Network measurement/management, Network virtualization, Network service, Blockchain, Security, Network intelligence/AI, etc. |
| Paper Information |
| Registration To |
NS |
| Conference Code |
2026-11-NS-ICM-NV-HCL |
| Language |
Japanese |
| Title (in Japanese) |
(See Japanese page) |
| Sub Title (in Japanese) |
(See Japanese page) |
| Title (in English) |
Wide-Area Load Balancing for LLM Inference in AI-and-RAN Environments Using Inter-Region KV Cache Migration and Pipeline Parallelism |
| Sub Title (in English) |
|
| Keyword(1) |
LLM Inference |
| Keyword(2) |
KV Cache |
| Keyword(3) |
Pipeline Parallelism |
| Keyword(4) |
AI-and-RAN |
| Keyword(5) |
|
| Keyword(6) |
|
| Keyword(7) |
|
| Keyword(8) |
|
| 1st Author's Name |
Taichi Kondo |
| 1st Author's Affiliation |
The University of Tokyo (The Univ. of Tokyo) |
| 2nd Author's Name |
Shota Ono |
| 2nd Author's Affiliation |
The University of Tokyo (The Univ. of Tokyo) |
| 3rd Author's Name |
Akihiro Nakao |
| 3rd Author's Affiliation |
The University of Tokyo (The Univ. of Tokyo) |
| 4th Author's Name |
|
| 4th Author's Affiliation |
() |
| 5th Author's Name |
|
| 5th Author's Affiliation |
() |
| 6th Author's Name |
|
| 6th Author's Affiliation |
() |
| 7th Author's Name |
|
| 7th Author's Affiliation |
() |
| 8th Author's Name |
|
| 8th Author's Affiliation |
() |
| 9th Author's Name |
|
| 9th Author's Affiliation |
() |
| 10th Author's Name |
|
| 10th Author's Affiliation |
() |
| 11th Author's Name |
|
| 11th Author's Affiliation |
() |
| 12th Author's Name |
|
| 12th Author's Affiliation |
() |
| 13th Author's Name |
|
| 13th Author's Affiliation |
() |
| 14th Author's Name |
|
| 14th Author's Affiliation |
() |
| 15th Author's Name |
|
| 15th Author's Affiliation |
() |
| 16th Author's Name |
|
| 16th Author's Affiliation |
() |
| 17th Author's Name |
|
| 17th Author's Affiliation |
() |
| 18th Author's Name |
|
| 18th Author's Affiliation |
() |
| 19th Author's Name |
|
| 19th Author's Affiliation |
() |
| 20th Author's Name |
|
| 20th Author's Affiliation |
() |
| 21st Author's Name |
|
| 21st Author's Affiliation |
() |
| 22nd Author's Name |
|
| 22nd Author's Affiliation |
() |
| 23rd Author's Name |
|
| 23rd Author's Affiliation |
() |
| 24th Author's Name |
|
| 24th Author's Affiliation |
() |
| 25th Author's Name |
|
| 25th Author's Affiliation |
() |
| 26th Author's Name |
/ / |
| 26th Author's Affiliation |
()
() |
| 27th Author's Name |
/ / |
| 27th Author's Affiliation |
()
() |
| 28th Author's Name |
/ / |
| 28th Author's Affiliation |
()
() |
| 29th Author's Name |
/ / |
| 29th Author's Affiliation |
()
() |
| 30th Author's Name |
/ / |
| 30th Author's Affiliation |
()
() |
| 31st Author's Name |
/ / |
| 31st Author's Affiliation |
()
() |
| 32nd Author's Name |
/ / |
| 32nd Author's Affiliation |
()
() |
| 33rd Author's Name |
/ / |
| 33rd Author's Affiliation |
()
() |
| 34th Author's Name |
/ / |
| 34th Author's Affiliation |
()
() |
| 35th Author's Name |
/ / |
| 35th Author's Affiliation |
()
() |
| 36th Author's Name |
/ / |
| 36th Author's Affiliation |
()
() |
| Speaker |
Author-1 |
| Date Time |
2026-11-05 10:00:00 |
| Presentation Time |
20 minutes |
| Registration for |
NS |
| Paper # |
NS2026-121 |
| Volume (vol) |
vol.126 |
| Number (no) |
no.222 |
| Page |
pp.7-12 |
| #Pages |
6 |
| Date of Issue |
2026-10-29 (NS) |