The AR/VR community agrees that ideal end-to-end latency for immersive communication should be 10–50 ms. Achieving this with current cloud computing and network technology is extremely difficult — round-trip times to centralized cloud data centers alone often exceed this budget.

LLRIS (Low-Latency for Augmented Reality Interactive Systems) explores an edge-cloud hybrid model where computation and cached results are placed near users, reducing latency for interactive AR/VR workloads. The project investigates:

  • Where to place computation — edge nodes, regional cloudlets, or centralized cloud
  • Content caching strategies — pre-fetching likely-needed AR objects near users
  • Network architecture — what network primitives best support sub-50ms interactive AR
  • Named Data Networking as transport — using NDN’s in-network caching for AR content delivery

This was a planning grant that laid groundwork for future large-scale AR/VR networking infrastructure proposals.

Funding: NSF CCRI Planning #2016714 ($132,200, TNTech share: $14,994, 2020–2022)

Team: Susmit Shannigrahi (Co-PI)