Layout-aware I/O Scheduling for terabits data movement

Abstract

Many science facilities, such as the Department of Energy's Leadership Computing Facilities and experimental facilities including the Spallation Neutron Source, Stanford Linear Accelerator Center, and Advanced Photon Source, produce massive amounts of experimental and simulation data. These data are often shared among the facilities and with collaborating institutions. Moving large datasets over the widearea network (WAN) is a major problem inhibiting collaboration. Next-generation, terabit-networks will help alleviate the problem, however, the parallel storage systems on the endsystem hosts at these institutions can become a bottleneck for terabit data movement. The parallel storage system (PFS) is shared by simulation systems, experimental systems, analysis and visualization clusters, in addition to wide-area data movers. These competing uses often induce temporary, but significant, I/O load imbalances on the storage system, which impact the performance of all the users. The problem is a serious concern because some resources are more expensive (e.g. super computers) or have time-critical deadlines (e.g. experimental data from a light source), but parallel file systems handle all requests fairly even if some storage servers are under heavy load. This paper investigates the problem of competing workloads accessing the parallel file system and how the performance of wide-area data movement can be improved in these environments. First, we study the I/O load imbalance problems using actual I/O performance data collected from the Spider storage system at the Oak Ridge Leadership Computing Facility. Second, we present I/O optimization solutions with layout-awareness on end-system hosts for bulk data movement. With our evaluation, we show that our I/O optimization techniques can avoid the I/O congested disk groups, improving storage I/O times on parallel storage systems for terabit data movement.

CRBF: Cross-Referencing Bloom-Filter-Based Data Integrity Verification Framework for Object-Based Big Data Transfer Systems

“…As a result, Figure 1c reveals that objects belonging to a single logical file are transferred in a non-sequential manner. Object-based big data transfer systems (OBDTS) can achieve higher data throughput than traditional big data transfer frameworks due to their enhanced scheduling algorithms [8][9][10].…”

“…Given a dataset and a false-positive probability, the total number of bits required for a data integrity solution can be calculated by substituting n in Equation (7).…”

“…However, similar to f Sum, D2 workloads have slightly higher memory requirements than the TPBF data integrity method, as the file size and stripe size are the same (i.e., 1 MB). Therefore, based on (7), the CRBF data integrity method exhibits a similar or higher memory overhead than that of the TPBF and f Sum solutions. Overall, Figure 10 illustrates that, similar to the TPBF, the proposed CRBF data integrity method is significant regarding memory savings when the dataset contains large or mixed-size files.…”

“…In a PFS (parallel file system), users share resources, which can lead to competition for those resources. This can cause a significant difference between the estimated and actual I/O performance [7]. Therefore, resources must be prevented from being congested during I/O operations to achieve the expected I/O performance.…”

“…Therefore, resources must be prevented from being congested during I/O operations to achieve the expected I/O performance. To achieve this objective, researchers proposed a layout-aware object-based bulk data movement framework (layout-aware data scheduling or LADS), optimizing I/O load imbalance issues in PFSs using layout-aware I/O scheduling and object storage target (OST) congestion-aware I/O scheduling algorithms [8][9][10]. Object-based big data transfer systems (OBDTS) help achieve higher end-to-end data transfer rates than state-of-the-art data transfer frameworks owing to these enhanced scheduling algorithms.…”