scispace - formally typeset
Search or ask a question

Showing papers by "Jose Renato Santos published in 2011"


14 Jun 2011
TL;DR: Experiments on the Xen-based implementation of Gatekeeper in a datacenter cluster demonstrate effective and flexible control of ingress/egress link bandwidth for tenant virtual machines under both TCP and greedy unresponsive UDP traffic.
Abstract: Cloud environments should provide network performance isolation for co-located untrusted tenants in a virtualized datacenter. We present key properties that a performance isolation solution should satisfy, and present our progress on Gatekeeper, a system designed to meet these requirements. Experiments on our Xen-based implementation of Gatekeeper in a datacenter cluster demonstrate effective and flexible control of ingress/egress link bandwidth for tenant virtual machines under both TCP and greedy unresponsive UDP traffic.

211 citations


Patent
07 Jun 2011
TL;DR: In this paper, the authors propose a method to determine if a packet's signature matches the stored packet signature, performing an action associated with the packet signature in order to perform packet processing for packets from virtual machines.
Abstract: Packet processing for packets from a virtual machine includes receiving a packet from an external switch at a computer system hosting a plurality of virtual machines. If the received packet is a learning packet, storing a packet signature determined from the learning packet. For a packet to be transmitted from a virtual machine in the computer system, determining if the packet's signature matches the stored packet signature. If the packet's signature matches the stored packet signature, performing an action associated with the packet signature.

56 citations


Patent
10 Feb 2011
TL;DR: In this paper, a hypervisor can scan local memory to determine the number of idle pages and a number of virtual machines, including a page map and a remote address destination for each idle page, by the hypervisor.
Abstract: Remote memory can be used for a number idle pages located on a virtual machine. A number of idle pages can be sent to the remote memory according to a placement policy, where the placement policy can include a number of weighting factors. A hypervisor on a computing device can record a local size and a remote page fault frequency of the number of virtual machines. The hypervisor can scan local memory to determine the number of idle pages and a number of idle virtual machines. The number of idle pages, including a page map and a remote address destination for each idle page, can be sent to the remote memory by the hypervisor. The number of virtual machines can be analyzed to determine a per-virtual machine local memory allocation.

28 citations


Patent
17 Oct 2011
TL;DR: In this article, non-congestion-controlled message traffic in a datacenter is managed by encapsulating the sending server and sending through a virtual circuit having a message transport protocol with rate control.
Abstract: Managing non-congestion-controlled message traffic in a datacenter. Any non-congestion-controlled packets originating in a sending server and intended for a receiving server are encapsulated in the sending server and transmitted through a virtual circuit having a message transport protocol with rate control. The packets are decapsulated in the receiving server.

7 citations


01 Jan 2011
TL;DR: A software-based prototype of memory disaggregation is built and examined how the additional level of indirection provided by the memory blade can provide significantly higher levels of consolidation.
Abstract: Recent architecture research has introduced a new building block - a memory blade - which provides disaggregated memory capacity expansion and sharing for an ensemble of blade servers. In this paper, we examine the systems implications of this new architectural building block. We build a software-based prototype of memory disaggregation and examine how the additional level of indirection provided by the memory blade can provide significantly higher levels of consolidation. We specifically examine the use case of multiple large memory virtual machines (VMs) consolidated onto a single server. We explore content based sharing strategies to maximize the utilization of both local and remote memory. Using an in-memory database workload, our results show significantly higher levels of consolidation versus baseline servers (twice as many VMs than a memory-constrained baseline, providing 47% higher throughput), cost-effective memory expansion (28% better performance-per-dollar versus a large memory baseline), and effective content based sharing (up to 40%).

6 citations