Optimizing Post-Copy Live Migration with System-Level Checkpoint Using Fabric-Attached Memory

Chou, Cheng‐Chun; Chen, Yuan; Reddy, Narasimha; Gratz, Paul V.

doi:10.1109/mchpc49590.2019.00010

Cited by 7 publications

(4 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The load balancer calls the process migration manager in traditional computing systems such as the cluster computing system. The status of the source and destination computing elements does not change in the process of implementing the process migration activities (Singh, G., & Singh, P., 2021;Chou, C. C., Chen, Y., Milojicic, D., Reddy, N., & Gratz, P., 2019, November). Therefore, the condition of matching the data structure of the migrant process to the corresponding data structure in the destination computing element be established.…”

Section: 𝐹𝐹(𝑙𝑙𝑙𝑙𝑙𝑙ℎ𝑖𝑖𝑖𝑖𝑖𝑖)mentioning

confidence: 99%

“…In traditional expandable computing systems, if the status of the computing element changes due to the implementation of local activities, if the data structure of the migrating process is not matched with the state of the corresponding data structure in the destination computing element, the migration condition is not established (Stoyanov, R., & Kollingbaum, M. J., 2018, June). In the process migration manager of flushing, the and destination computing elements does not change in the process of implementing the process migration activities (Singh, G., & Singh, P., 2021;Chou, C. C., Chen, Y., Milojicic, D., Reddy, N., & Gratz, P., 2019, November). Therefore, the condition of matching the data structure of the migrant process to the corresponding data structure in the destination computing element be established.…”

Section: 𝐹𝐹(𝑙𝑙𝑙𝑙𝑙𝑙ℎ𝑖𝑖𝑖𝑖𝑖𝑖)mentioning

confidence: 99%

See 1 more Smart Citation

A Mechanism for Using the Flushing Process Migration Mechanism in Distributed Exascale Systems

Gholamrezaie¹,

Feyziyev²

2022

AzJHPC

View full text Add to dashboard Cite

The effect of dynamic and interactive events on the function of the elements that make up the computing system manager causes the time required to run the user program to increase or the operation of these elements to change. These changes either increase the execution time of the scientific program or make the system incapable of executing the program. Computational processes on the migration process and vector algebras try to analyze and enable the Flushing process migration mechanism in support of distributed Exascale systems despite dynamic and interactive events. This paper investigates the Flushing process migration management mechanism in distributed Exascale systems, the effects of dynamic and interactive occurrences in the computational system, and the impact of dynamic and interactive events on the system.

show abstract

Section: 𝐹𝐹(𝑙𝑙𝑙𝑙𝑙𝑙ℎ𝑖𝑖𝑖𝑖𝑖𝑖)mentioning

confidence: 99%

Section: 𝐹𝐹(𝑙𝑙𝑙𝑙𝑙𝑙ℎ𝑖𝑖𝑖𝑖𝑖𝑖)mentioning

confidence: 99%

A Mechanism for Using the Flushing Process Migration Mechanism in Distributed Exascale Systems

Gholamrezaie¹,

Feyziyev²

2022

AzJHPC

View full text Add to dashboard Cite

show abstract

“…MOSIQS's target architecture is an array of PM devices distributed across hundreds of compute nodes. PM on each compute node is shared with other compute nodes via a shared PM pool abstraction via highspeed fabric attached memory (FAM) interconnect such as Gen-Z [6], [45]. Figure 2 depicts a high-level architectural overview of the MOSIQS.…”

Section: B Target Architecturementioning

confidence: 99%

MOSIQS: Persistent Memory Object Storage With Metadata Indexing and Querying for Scientific Computing

et al. 2021

View full text Add to dashboard Cite

Scientific applications often require high-bandwidth shared storage to perform joint simulations and collaborative data analytics. Shared memory pools provide a chance to satisfy such needs. Recently, a high-speed network such as Gen-Z utilizing persistent memory (PM) offers an opportunity to create a shared memory pool connected to compute nodes. However, there are several challenges to use scientific applications on the shared memory pool directly such as scalability, failure-atomicity, and lack of scientific metadata-based search and query. In this paper, we propose MOSIQS, a persistent memory object storage framework with metadata indexing and querying for scientific computing. We design MOSIQS based on the key idea that memory objects on PM pool can live beyond the application lifetime and can become the sharing currency for applications and scientists. MOSIQS provides an aggregate memory pool atop an array of persistent memory devices to store and access memory objects to accelerate scientific computing. MOSIQS uses a lightweight persistent memory key-value store to manage the metadata of memory objects, which enables memory object sharing. To facilitate metadata search and query over millions of memory objects resident on memory pool, we introduce Group Split and Merge (GSM), a novel persistent index data structure designed primarily for scientific datasets. GSM splits and merges dynamically to minimize the query search space and maintains low query processing time while overcoming the index storage overhead. MOSIQS is implemented atop of PMDK. We evaluate the proposed approach on many-core server with an array of real PM devices. Experimental results show that MOSIQS gains a 100% write performance improvement and executes multi-attribute queries efficiently with 2.7× less index storage overhead offering significant potential to speed up scientific computing applications.

show abstract

“…This algorithm reduced the total migration time but the downtime was almost the same with the original post-copy. Chou et al [78] proposed a hardware approach using high-performance Fabric-Attached Memory (FAM) to interconnect physical machines. Memory pages are dumped from source node to FAM during the migration, while the VM is running at destination node.…”

Section: Robust Migration Related Workmentioning

confidence: 99%

Choice With No Regret – Mobility Induced Migration in Fog Computing

Talebian

View full text Add to dashboard Cite

Fog computing is a highly virtualized platform that acts as an intermediary be- tween end-devices and cloud centers. Due to the close proximity between fog nodes and users, fog computing is ideal for latency-sensitive applications. To ensure Quality of Service (QoS) for mobile users, especially in applications like autonomous vehicles, fog computing employs a network slicing approach for efficient resource allocation. Moreover, services need to be kept close to users, necessitating the transfer of services from current fog nodes to closer nodes while users are mobile, a process known as ”mobility induced migration in fog computing”.This thesis introduces three novel methods: (1) Fuzzy Migration Decision (FMD)), (2) Efficient Destination Choice (EDC) and (3) Robust Intra-Slice Migration in Fog Computing. Each of these methods plays a distinct role in the migration process, encompassing the decision-making phase of when to initiate migration, the selection of the optimal destination node within the network, and the execution of migration in a manner that prioritizes robustness and low latency.The FMD algorithm consistently outperforms existing methods by determining the optimal time to start migration using fuzzy logic. It achieves approximately a 15% reduction in migration time and a 12% decrease in network usage compared to bench- mark algorithms.Similarly, the EDC algorithm, which aims to find the optimal destination node us- ing fuzzy logic and avoids negative impacts on network performance caused by mi- gration, achieves a 10.15% reduction in network usage compared to benchmark algo- rithms.Moreover, our proposed Robust Intra-Slice Migration in Fog Computing algorithm demonstrates a reduction of total migration time by around 10-26% and downtime by about 2-23% compared to non-robust post-copy migration techniques.The evaluation results demonstrate that the proposed algorithms effectively find a trade off between preserving QoS for users and enhancing network efficiency dur- ing their operation. Furthermore, when we execute the entire migration process using the proposed algorithms, there is a notable reduction in both migration latency and the number of migrations. This migration process exhibits robust performance when compared to other studies. These findings underscore how addressing all three mi- gration challenges (When, Where, and How) can significantly influence the overall outcome of mobility-induced migration.

show abstract

Optimizing Post-Copy Live Migration with System-Level Checkpoint Using Fabric-Attached Memory

Cited by 7 publications

References 16 publications

A Mechanism for Using the Flushing Process Migration Mechanism in Distributed Exascale Systems

A Mechanism for Using the Flushing Process Migration Mechanism in Distributed Exascale Systems

MOSIQS: Persistent Memory Object Storage With Metadata Indexing and Querying for Scientific Computing

Choice With No Regret – Mobility Induced Migration in Fog Computing

Contact Info

Product

Resources

About