Advanced data centers strive for high performance and throughput, which can be achieved through the desirable merits of Remote Procedure Call (RPC) programming model and the low latency of Remote Direct Memory Access (RDMA). However, despite the widespread availability of these software and hardware utilities, they have been utilized separately for their own applications in existing production systems for many years. Although researchers have attempted to develop RDMA-enabled RPC prototypes, they often face challenges such as API discrepancies and a lack of specific features for effective integration with major production software, rendering them incompatible. This industry R&D project aims to enhance the performance of gRPC, a widely utilized RPC framework in major companies, by integrating RDMA as an internal component. Our system solution, called, combines the simple user interface and other merits of gRPC with low latency for remote data accesses. RR-Compound is fully compatible with gRPC and can serve as a seamless replacement without altering existing applications. However, to achieve low latency, high throughput, and scalability for RR-Compound, several technical challenges in managing network connections and memory space utilization must be effectively addressed. To overcome the limitations of existing connection methods, we have developed a new method called BPEV that is independent of gRPC and applicable to all RDMA systems. We have also retained the asynchronous framework of gRPC, albeit with limited buffer space in RDMA memory management. In micro-benchmarks, RR-Compound outperforms mRPC - the state-of-the-art RPC framework for a large number of connections, achieving a 14.77% increase in throughput and a 42.55% reduction in latency. Subsequently, we compare RR-Compound with gRPC over IPoIB using two real-world applications: KV-Store and TensorFlow. RR-Compound achieves up to a 2.35x increase in throughput and reduces the average latency by 46.92%.

The popular microservice architecture for applications brings new challenges for consistent distributed transactions across multiple microservices. These microservices may be implemented in different languages, and access multiple underlying databases. Consistent distributed transactions are a real requirement but are very hard to achieve with existing technologies in these environments. In this demo we present GRIT: a system that resolves this challenge by cleverly leveraging deterministic database technologies and optimistic concurrency control protocol(OCC). A transaction is optimistically executed with its read-set and write-set captured during the execution phase. Then at the commit time, conflict checking is performed and a global commit decision is made. A logically committed transaction is persisted into logs first, and then asynchronously applied to the physical databases deterministically. GRIT is able to achieve consistent, high throughput and serializable distributed transactions for any applications invoking microservices. The demonstration offers a walk-through of how GRIT can easily support distributed transactions across multiple microservices and databases.

We demonstrate Grano 1 , an end-to-end anomaly detection and root cause analysis (or RCA for short) system for cloud-native distributed data platform by providing a holistic view of the system component topology, alarms and application events. Grano provides: a Detection Layer to process large amount of time-series monitoring data to detect anomalies at logical and physical system components; an Anomaly Graph Layer with novel graph modeling and algorithms for leveraging system topology data and detection results to identify the root cause relevance at the system component level; and an Application Layer that automatically notifies on-call personnel and presents real-time and on-demand RCA support through an interactive graph interface. The system is deployed and evaluated using eBay's production data to help on-call personnel to shorten the identification of root cause from hours to minutes.

The sheer growth in data volume and Hadoop cluster size make it a significant challenge to diagnose and locate problems in a production-level cluster environment efficiently and within a short period of time. Often times, the distributed monitoring systems are not capable of detecting a problem well in advance when a large-scale Hadoop cluster starts to deteriorate i n performance or becomes unavailable. Thus, inc o m i n g workloads, scheduled between the time when cluster starts to deteriorate and the time when the problem is identified, suffer from longer execution times. As a result, both reliability and throughput of the cluster reduce significantly. In this paper, we address this problem by proposing a system called Astro, which consists of a predictive model and an extension to the Hadoop scheduler. The predictive model in Astro takes into account a rich set of cluster behavioral information that are collected by monitoring processes and model them using machine learning algorithms to predict future behavior of the cluster. The Astro predictive model detects anomalies in the cluster and also identifies a ranked set of metrics that have contributed the most towards the problem. The Astro scheduler uses the prediction outcome and the list of metrics to decide whether it needs to move and reduce workloads from the problematic cluster nodes or to prevent additional workload allocations to them, in order to improve both throughput and reliability of the cluster. The results demonstrate that the Astro scheduler improves usage of cluster compute resources significantly by 64.23% compared to traditional Hadoop. Furthermore, the runtime of the benchmark application reduced by 26.68% during the time of anomaly, thus improving the cluster throughput.