A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL.
The file type is application/pdf
.
Filters
One way to offer consistent services on such unreliable systems are replicated state machines (RSMs). ...
Our protocols enable operators to fully utilize the new capabilities of the quickly growing number of RDMA-capable datacenter networks. ...
shepherd Jay Lofstead, the anonymous reviewers; Miguel Castro for his insightful feedback; Timo Schneider for helpful discussions; and the Systems Group at ETH Zurich for providing us the InfiniBand machine ...
doi:10.1145/2749246.2749267
dblp:conf/hpdc/PokeH15
fatcat:vpvxg7sg3jhytmndiuqoen5kh4
State machine replication (SMR) uses Paxos to enforce the same inputs for a program (e.g., Redis) replicated on a number of hosts, tolerating various types of failures. ...
Unfortunately, traditional Paxos protocols incur prohibitive performance overhead on server programs due to their high consensus latency on TCP/IP. ...
INTRODUCTION State machine replication (SMR) runs the same program on replicas of hosts and invokes a distributed consensus protocol (typically, Paxos [53] ) to enforce the same total order of inputs ...
doi:10.1145/3127479.3128609
dblp:conf/cloud/WangJCYC17
fatcat:y6en5cvguraiboak47td3okptm
Towards Low-Latency Byzantine Agreement Protocols Using RDMA
2018
2018 48th Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshops (DSN-W)
Retaining the performance benefits for RDMA during its integration, however, is non-trivial and error-prone. ...
For this, we present RUBIN, a framework offering similar functionality to the Java NIO selector, which can handle multiple network connections efficiently with a single thread and is employed in several ...
DARE [7] , an RDMA-tailored replicated state machine protocol, aims to optimize for low latency in replica communication. ...
doi:10.1109/dsn-w.2018.00054
dblp:conf/dsn/RuschMK18
fatcat:olcgouoit5go7kree62flpyblu
Datacenter RPCs can be General and Fast
[article]
2019
arXiv
pre-print
It is commonly believed that datacenter networking software must sacrifice generality to attain high performance. ...
We port a production-grade implementation of Raft state machine replication to eRPC without modifying the core Raft source code. ...
Codesign with RDMA is popular, with numerous examples from key-value stores [25, 38, 50, 65, 66] , state machine replication [58] , and transaction processing systems [21, 26, 41, 66] . ...
arXiv:1806.00680v2
fatcat:ryqqflbh4rgqpf2voguhgdhs6e
Recent Advancements In Distributed System Communications
[article]
2021
arXiv
pre-print
These issues degrade the performance of distributed systems, which rely on fast communications between machines to be able to serve a large number of client requests with low latency and high throughput ...
The former require specialized hardware to offer high communications performance. ...
high performance. ...
arXiv:2107.01381v1
fatcat:vr4cdreyv5ho5li4hpygmxmcbu
Cyclone: High Availability for Persistent Key Value Stores
[article]
2017
arXiv
pre-print
Unfortunately their performance is hampered by the need to maintain and replicate a write ahead log to guarantee availability in the face of machine and storage failures. ...
replication. ...
DARE [30] implements state machine replication over RDMA, using a custom replication protocol. DARE does not consider persistence in its design. ...
arXiv:1711.06964v1
fatcat:an5plvrycbbpld627vguvqftey
Invalidation-Based Protocols for Replicated Datastores
[article]
2021
arXiv
pre-print
Based on this insight, we draw inspiration from the multiprocessor for high-performance, strongly consistent replication in the datacenter. ...
The multiprocessor regime uses invalidations to afford strongly consistent replication with high performance but neglects fault tolerance. ...
(Accessed on
26/07/2021).
[181] Marius Poke and Torsten Hoefler. DARE: High-Performance State Ma-
chine Replication on RDMA Networks. ...
arXiv:2112.02405v1
fatcat:toi3wflouvbzngpdfzypw3l4li
Fiber-based architecture for NFV cloud databases
2017
Proceedings of the VLDB Endowment
This transformation is referred to as Network Function Virtualization (NFV). ...
The telco industry is gradually shifting from using monolithic software packages deployed on custom hardware to using modular virtualized software functions deployed on cloudified data centers using commodity ...
On the other hand, as operators dare to upgrade or build modern data centers, deployments may need also to scale-up on physical machines with more cores. ...
doi:10.14778/3137765.3137774
fatcat:tlzqna6ukzdy7f7mfqz3qtsipu
Designing the replication layer of a general-purpose datacenter key-value store
[article]
2022
We start the exploration by observing that modern, server-grade hardware with manycore servers and RDMA-capable networks, challenges conventional wisdom in protocol design. ...
We focus on strong consistency semantics because they are necessary for a general-purpose KVS and they are at odds with performance. ...
RDMA-capable networking. ...
doi:10.7488/era/1772
fatcat:pmecynsjqzfbdcsw44k2dxytk4
Invalidation-based protocols for replicated datastores
[article]
2021
Based on this insight, we draw inspiration from the multiprocessor for high-performance, strongly consistent replication in the datacenter. ...
The multiprocessor regime uses invalidations to afford strongly consistent replication with high performance but neglects fault tolerance. ...
(Accessed on
26/07/2021).
[181] Marius Poke and Torsten Hoefler. DARE: High-Performance State Ma-
chine Replication on RDMA Networks. ...
doi:10.7488/era/1639
fatcat:snjzrhkqqnc7dpsfe4gzaskhqa
Programming Safely with Weak (and Strong) Consistency
2020
DARE [Poke and Hoefler, 2015]
looks at state machine replication on an RDMA network. ...
Our experiments seek to answer the following questions: • How do the core state machine replication protocols perform on modern RDMA hardware, and on data-center TCP running over 100Gbps Ethernet? ...
doi:10.7298/4pje-j388
fatcat:3o2ydxfvp5e4fgus3z6i7zxb64
Dagstuhl Reports, Volume 5, Issue 10, October 2015, Complete Issue
[article]
2016
We will present a cyclically time-expanded network model for simultaneously optimizing traffic assignment and traffic signals in an urban road network with a special focus on user equilibria. ...
In future scenarios for transport in Smart Cities, mobility providers might control the routes taken by a significant share of vehicles on the road. ...
State machine replication as a service. The hardware can provide these services to make it easier for the OS mechanisms: Warning before failure. ...
doi:10.4230/dagrep.5.10
fatcat:u44u5pkszbayti24uerfh3wifu
Death sentences : the modern ordering of mortality
2005
This transformation (or tension, in my view) is underscored by the discovery of an interconnection between the way people die, on the one hand, and political order and economic prosperity, on the other ...
the symbolic significance of death to one that treats death discursively as a medical and public health problem. ...
( Automaton) is an imitation of the natural machine (the human body), so is the state (the Leviathan) an artificial machine made on the model of the human body. ...
doi:10.14288/1.0092768
fatcat:zjazk5cimba33groitylm6c7s4