ParaCell: 基于轻量级容器内隔离和意图驱动内存管理的虚拟化安全容器
ParaCell: Paravirtualized Secure Containers with Lightweight Intra-Container Isolation and Intent-Driven Memory Management
AI总结:
ParaCell通过引入轻量级容器内隔离和意图驱动内存管理,解决了传统容器系统中隔离与性能之间的根本性权衡问题,提升了容器运行效率和内存弹性。
摘要:
安全容器通过每个容器拥有自己的内核来隔离容器,缓解了传统容器系统中普遍存在的共享内核攻击问题。然而,现有设计仍面临一个根本性的隔离-性能权衡问题。云部署嵌套放大了虚拟机退出和页面表管理的成本,而新兴的代理工作负载暴露了突发性的内存需求,需要细粒度的弹性。本文认为这一权衡源于两个根本原因。首先,现有设计缺乏轻量级容器内隔离原语,用于频繁的容器用户-内核切换。其次,主机将容器内存管理视为不透明,迫使反应式二次故障和粗粒度的大页面映射来摊销其成本。本文提出了ParaCell,一种基于两个洞察的虚拟化安全容器运行时。首先,内存地址空间硬件保护原语可以提供轻量级的容器内隔离。ParaCell使用基于MPK的XGates来隔离容器用户和容器内核,使其在单一地址空间内,将频繁的用户-内核切换转换为直接的域切换。其次,容器内核分配器已经编码了内存管理的意图。ParaCell引入Pager来拦截分配和释放事件,批量主动将GPA到HPA的绑定和解除绑定,并避免反应式影子页面表故障,同时保持细粒度的内存弹性。ParaCell作为RunV的直接替换实现。我们的实验表明,在传统云和新兴代理应用中,ParaCell在裸金属和嵌套设置下,相比PVM减少了高达57%和79%的延迟,相比RunV减少了高达33%和88%的延迟。在代理工作负载上,ParaCell相比最先进的虚拟机内存回收技术HyperAlloc节省了高达35.6%的内存。
英文摘要:
Secure containers isolate each container with its own kernel, mitigating shared-kernel attacks prevalent in traditional container systems. However, existing designs still face a fundamental isolation--performance trade-off. Nested-cloud deployments amplify the cost of VM exits and page-table management, while emerging agentic workloads expose bursty memory demand that requires fine-grained elasticity. We attribute this trade-off to two root causes. First, existing designs lack lightweight intra-container isolation primitives for frequent container user--kernel transitions. Second, the host treats container memory management as opaque, forcing reactive secondary faults and coarse-grained huge page mappings to amortize their cost. This paper presents ParaCell, a paravirtualized secure container runtime built on two insights. First, intra-address-space hardware protection primitives can provide lightweight intra-container isolation. ParaCell uses MPK-based XGates to isolate the container user and container kernel within a single address space, turning frequent user--kernel transitions into direct domain switches. Second, container kernel allocators already encode memory-management intent. ParaCell introduces Pager to interpose on allocation and free events, batch proactive GPA to HPA bindings and unbindings, and avoid reactive shadow page-table faults while preserving fine-grained memory elasticity. ParaCell is implemented as a drop-in replacement for RunV. Our experiments demonstrate that, across traditional cloud and emerging agent applications, ParaCell reduces latency by up to 57% and 79% over PVM, and by up to 33% and 88% over RunV, in bare-metal and nested setups, respectively. On agent workloads, ParaCell saves up to 35.6% memory compared with the state-of-the-art VM memory reclamation technique, HyperAlloc.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。