Overlay networks serve as the de facto network virtualization technique for providing connectivity among distributed containers. Despite the flexibility in building customized private container networks, overlay networks incur significant performance loss compared to physical networks (i.e., the native). The culprit lies in the inclusion of multiple network processing stages in overlay networks, which prolongs the network processing path and overloads CPU cores. In this paper, we propose mFlow, a novel packet steering approach to parallelize the in-kernel data path of network flows. mFlow exploits packet-level parallelism in the kernel network stack by splitting the packets of the same flow into multiple micro-flows, which can be processed in parallel on multiple cores. mFlow devises new, generic mechanisms for flow splitting while preserving in-order packet delivery with little overhead. Our evaluation with both micro-benchmarks and real-world applications demonstrates the effectiveness of mFlow, with significantly improved performance – e.g., by 81% in TCP throughput and 139% in UDP compared to vanilla overlay networks. mFlow even achieved higher TCP throughput than the native (e.g., 29.8 vs. 26.6 Gbps).
more »
« less
Product Kernel Interpolation for Scalable Gaussian Processes
Recent work shows that inference for Gaussian processes can be performed efficiently using iterative methods that rely only on matrix-vector multiplications (MVMs). Structured Kernel Interpolation (SKI) exploits these techniques by deriving approximate kernels with very fast MVMs. Unfortunately, such strategies suffer badly from the curse of dimensionality. We develop a new technique for MVM based learning that exploits product kernel structure. We demonstrate that this technique is broadly applicable, resulting in linear rather than exponential runtime with dimension for SKI, as well as state-of-the-art asymptotic complexity for multi-task GPs.
more »
« less
- Award ID(s):
- 1525919
- NSF-PAR ID:
- 10065065
- Date Published:
- Journal Name:
- AISTATS 2018
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Modern operating systems are monolithic. Today, however, lack of isolation is one of the main factors undermining security of the kernel. Inherent complexity of the kernel code and rapid development pace combined with the use of unsafe, low-level programming language results in a steady stream of errors. Even after decades of efforts to make commodity kernels more secure, i.e., development of numerous static and dynamic approaches aimed to prevent exploitation of most common errors, several hundreds of serious kernel vulnerabilities are reported every year. Unfortunately, in a monolithic kernel a single exploitable vulnerability potentially provides an attacker with access to the entire kernel.Modern kernels need isolation as a practical means of confining the effects of exploits to individual kernel subsystems. Historically, introducing isolation in the kernel is hard. First, commodity hardware interfaces provide no support for efficient, fine-grained isolation. Second, the complexity of a modern kernel prevents a naive decomposition effort. Our work on Lightweight Execution Domains (LXDs) takes a step towards enabling isolation in a full-featured operating system kernel. LXDs allow one to take an existing kernel subsystem and run it inside an isolated domain with minimal or no modifications and with a minimal overhead. We evaluate our approach by developing isolated versions of several performance-critical device drivers in the Linux kernel.more » « less
-
Subteam Replacement: given a team of people em- bedded in a social network to complete a certain task, and a subset of members (i.e., subteam) in this team which have become unavailable, find another set of people who can perform the subteam’s role in the larger team. We conjecture that a good candidate subteam should have high skill and structural similarity with the replaced subteam while sharing a similar connection with the larger team as a whole. Based on this conjecture, we propose a novel graph kernel which evaluates the goodness of candidate subteams in this holistic way freely adjustable to the need of the situation. To tackle the significant computational difficulties, we equip our kernel with a fast approximation algorithm which (a) employs effective pruning strategies, (b) exploits the similarity between candidate team structures to reduce kernel computations, and (c) features a solid theoretical bound on the quality of the obtained solution. We extensively test our solution on both synthetic and real datasets to demonstrate its effectiveness and efficiency. Our proposed graph kernel outputs more human-agreeable recommendations compared to metrics used in previous work, and our algorithm consistently outperforms alternative choices by finding near- optimal solutions while scaling linearly with the size of the replaced subteam.more » « less
-
more » « less
Abstract Ultra‐high molecular weight polyethylene (UHMWPE) is one of the most prominent high‐performance thermoplastics for biomedical, leisure, and coating applications. Large‐scale recycling of UHMWPE is extremely difficult due to the high melt viscosity of the material as well as its exceptional chemical resistance and impact strength. There is a need for a commercially scalable methodology that can process the waste feedstock for mechanical recycling while sustaining the outstanding physical properties of the material. Solid‐state shear pulverization (SSSP) is a continuous, twin‐screw extruder‐based processing technique in which the low‐temperature application of shear and compressive forces impart changes in structure at different length scales to overcome the challenges of difficult‐to‐recycle polymers. This paper investigates the use of SSSP in mechanically recycling post‐industrial scrap UHMWPE (rUHMWPE) material from a local ski and snowboard manufacturer. The SSSP‐processed particles are flat, micron‐scale flakes with enhanced surface area, which can sinter very quickly when compression molded. The molded rUHMWPE samples in turn exhibit enhanced ductility and toughness compared to the as‐received scrap material, based on the tunable mechanochemical modification of the ethylene chains.
-
Fast, byte-addressable persistent memory (PM) is becoming a reality in products. However, porting legacy kernel file systems to fully support PM requires substantial effort and encounters the challenge of bridging the gap between block-based access granularity and byte-addressability. Moreover, new PM-specific file systems remain far from production-ready, preventing them from being widely used. In this paper, we propose P2CACHE, a novel in-kernel caching mechanism to explore how legacy kernel file systems can effectively evolve in the face of fast, byte-addressable PM. P2CACHE exploits a read/write-distinguishable memory hierarchy upon a tiered memory system involving both PM and DRAM. P2CACHE leverages PM to serve all write requests for instant data durability and strong crash consistency while using DRAM to serve most read I/Os for high I/O performance. Further, P2CACHE employs a simple yet effective synchronization model between PM and DRAM by leveraging device-level parallelism. Our evaluation shows that P2CACHE can significantly increase the performance of legacy kernel file systems -- e.g., by 200x for RocksDB on Ext4 -- meanwhile equipping them with instant data durability and strong crash consistency, similar to PM-specialized file systems.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
