With the growing importance of in-memory data processing, cloud service providers have launched large memory virtual machine services to accommodate memory intensive workloads. Such large memory services using low volume scaled-up machines are far less cost-efficient than scaled-out services consisting of high volume commodity servers. By exploiting memory usage imbalance across cloud nodes, disaggregated memory can scale up the memory capacity for a virtual machine in a cost-effective way. Disaggregated memory allows available memory in remote nodes to be used for the virtual machine requiring more memory than its locally available memory. It supports high performance with the faster direct memory while satisfying the memory capacity demand with the slower remote memory. This paper proposes a new hypervisor-integrated disaggregated memory system for cloud computing. The hypervisor-integrated design has several new contributions in its disaggregated memory design and implementation. First, with the tight hypervisor integration, it investigates a new page management mechanism and policy tuned for disaggregated memory in virtualized systems. Second, it restructures the memory management procedures and relieves the scalability concern for supporting large virtual machines. Third, exploiting page access records available to the hypervisor, it supports application-aware elastic block sizes for fetching remote memory pages with different granularities. Depending on the degrees of spatial locality for different regions of memory in a virtual machine, the optimal block size for each memory region is dynamically selected. The experimental results with the implementation integrated to the KVM hypervisor, show that the disaggregated memory can provide on average 6 percent performance degradation compared to the ideal local-memory only machine, even though the direct memory capacity is only 50 percent of the total memory footprint.

Due to increased demand for high scalability and programmability in recent data centers, there has been a paradigm shift from traditional hardware-based network functions to software defined network functions. While implementing network functions in software offers great flexibility, it also reveals significant overheads and performance issues associated with the network stack in current operating systems. Since manycore processors and advanced network technology can provide significant performance improvements for applications such as router workloads that are responsive to parallelism, there is a need to optimize the packet processing performance of current operating systems. The network stack in the operating systems suffer from a number of overheads associated with core mechanisms of the kernel such as task scheduling, memory management, process synchronization, interrupt handling, etc. Each of these overheads snowballs into severe performance penalty of software network functions as each packet traverses its path through the network stack in the kernel. In this paper, we provide an analysis of the packet processing capability of manycore systems, running general-purpose operating system (OS). We identify the major performance bottlenecks in the packet processing path of Linux kernel network stack and provide an insight into various performance and scalability issues that need to be addressed.