Streaming network traffic measurement and analysis is critical for detecting and preventing any real-time anomalies in the network. The high speeds and complexity of today's networks, coupled with ever evolving threats, necessitate closing of the loop between measurements and their analysis in real time. The ensuing system demands high levels of programmability and processing where streaming measurements adapt to the changing network behavior in a goal-oriented manner. In this work, we exploit the features and requirements of the problem and develop an application-specific FPGA-based closed-loop measurement (CLM) system. We make novel use of fine-grained partial dynamic reconfiguration (PDR) as underlying reprogramming paradigm, performing low-latency just-in-time compiled logic changes in FPGA fabric corresponding to the dynamic measurement requirements. Our innovative dynamically reconfigurable socket offers 3× logic savings over conventional static solutions, while offering much reduced reconfiguration latencies over conventional PDR mechanisms. We integrate multiple sockets in a highly parallel CLM framework and demonstrate its effectiveness in identifying heavy flows in streaming network traffic. The results using an FPGA prototype offer 100 percent detection accuracy while sustaining increasing link speeds. © 1968-2012 IEEE.

Online network traffic measurements and analysis is critical for detecting and preventing any real-time anomalies in the network. We propose, implement, and evaluate an online, adaptive measurement platform, which utilizes real-time traffic analysis results to refine subsequent traffic measurements. Central to our solution is the concept of Multi-Resolution Tiling (MRT), a heuristic approach that performs sequential analysis of traffic data to zoom into traffic subregions of interest. However, MRT is sensitive to transient traffic spikes. In this paper, we propose three novel traffic streaming algorithms that overcome the limitations of MRT and can cater to varying degrees of computational and storage budgets, detection latency, and accuracy of query response. We evaluate our streaming algorithms on a highly parallel and programmable hardware as well as a traditional software-based platforms. The algorithms demonstrate significant accuracy improvement over MRT in detecting anomalies consisting of synthetic hard-to-track elephant flows and global icebergs. Our proposed algorithms maintain the worst-case complexities of the MRT while incurring only a moderate increase in average resource utilization. © 2013 IEEE.

Background

Allogeneic hematopoietic cell transplantation (HCT) can be curative for many hematologic diseases. However, complications such as graft-versus-host disease (GVHD) and relapse of primary malignancy remain significant and are the leading causes of morbidity and mortality. Effects of killer Ig-like receptors (KIR)-influenced NK cells on HCT outcomes have been extensively pursued over the last decade. However, the relevance of the reported algorithms on HLA matched myeloablative HCT with rabbit antithymocyte globulin (ATG) is used for GVHD prophylaxis remains elusive. Here we examined the role of KIR and KIR-ligands of donor-recipient pairs in modifying the outcomes of ATG conditioned HLA matched sibling and unrelated donor HCT.

Methods and findings

The study cohort consisted of 281 HLA matched sibling and unrelated donor-recipient pairs of first allogeneic marrow or blood stem cell transplantation allocated into 'discovery' (135 pairs) and 'validation' (146 pairs) cohorts. High resolution HLA typing was obtained from the medical charts and KIR gene repertoires were obtained by a Luminex® based SSO method. All surviving patients were followed-up for a minimum of two years. KIR and HLA class I distributions of HCT pairs were stratified as per applicable definitions and were tested for their association with cause specific outcomes [acute GVHD grade II-IV (aGVHD), chronic GVHD needing systemic therapy (cGVHD) and relapse] using a multivariate competing risks regression model as well as with survival outcomes [relapse-free survival (RFS), cGVHD & relapse free survival (cGRFS) and overall survival (OS)] by multivariate Cox proportional hazards regression model. A significant association between KIR genotype mismatching (KIR-B/x donor into KIR-AA recipient or vice versa) and cGVHD was found in both discovery (p = 0.001; SHR = 2.78; 95%CI: 1.50-5.17) and validation cohorts (p = 0.005; SHR = 2.61; 95%CI: 1.33-5.11). High incidence of cGVHD associated with KIR genotype mismatching was applicable to both sibling and unrelated donors and was specific to recipients who had one or two C1 bearing HLA-C epitopes (HLA-C1/x, p = 0.001; SHR = 2.40; 95%CI: 1.42-4.06). When compared with KIR genotype mismatched transplants, HLA-C1/x patients receiving grafts from KIR genotype matched donors had a significantly improved cGRFS (p = 0.013; HR = 1.62; 95%CI: 1.11-2.39). Although there was no effect of KIR genotype matching on survival outcomes, a significantly reduced incidence of relapse (p = 0.001; SHR = 0.22; 95%CI: 0.10-0.54) and improved relapse-free survival (p = 0.038; HR = 0.40; 95%CI: 0.17-0.95) was observed with one or more missing ligands for donor inhibitory KIR among the recipients of unrelated donor transplants.

Conclusions

The present study for the first time presents the beneficial effects of KIR genotype matching in reducing cGVHD in myeloablative transplant setting using HLA matched (sibling and unrelated) donors. The findings offer a clinically applicable donor selection strategy that can help control cGVHD without affecting the risk of relapse and/or identify patients at a high risk of developing cGVHD as potential candidates for preemptive therapy. The findings also affirm the beneficial effect of one or more missing inhibitory KIR ligands in the recipient in reducing relapse and improving a relapse free survival in unrelated donor transplants.