Master’s Thesis Presentation • Systems and Networking — Achieving Resource-Efficient Consensus with RDMAExport this event to calendar

Thursday, April 18, 2019 — 3:00 PM EDT

Mikhail Kazhamiaka, Master’s candidate
David R. Cheriton School of Computer Science

Sift is a new consensus protocol for replicating state machines. It disaggregates CPU and memory consumption by creating a novel system architecture enabled by one-sided RDMA operations. 

We show that this system architecture allows us to develop a consensus protocol design which centralizes the replication logic. This simplifies the design of the protocol by preventing complex interactions between the participants of the consensus group. Our dissaggregated design also enables Sift to reduce deployment costs by sharing backup computational nodes across consensus groups deployed within the same cloud environment. We can further reduce the required storage resources by integrating erasure codes without making significant changes to our protocol. 

Our evaluation results show that in a cloud environment with 100 groups where each group can support up to 2 simultaneous failures, Sift can reduce the cost by 56% compared to an RDMA-based Raft deployment.

Location 
DC - William G. Davis Computer Research Centre
2310
200 University Avenue West

Waterloo, ON N2L 3G1
Canada

S M T W T F S
29
30
31
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
1
  1. 2020 (22)
    1. March (1)
    2. February (1)
    3. January (20)
  2. 2019 (255)
    1. December (21)
    2. November (25)
    3. October (16)
    4. September (20)
    5. August (18)
    6. July (12)
    7. June (23)
    8. May (23)
    9. April (32)
    10. March (25)
    11. February (16)
    12. January (24)
  3. 2018 (220)
  4. 2017 (36)
  5. 2016 (21)
  6. 2015 (36)
  7. 2014 (33)
  8. 2013 (23)
  9. 2012 (4)
  10. 2011 (1)
  11. 2010 (1)
  12. 2009 (1)
  13. 2008 (1)