An official website of the United States government
Large cloud providers including AWS, Azure, and Google Cloud offer two tiers of network services to their customers: one class uses the providers' private wide area networks (WAN-transit) to carry a customer's traffic as much as possible, and the other uses the public internet (inet-transit). Little is known about how each cloud provider configures its network to offer different transit services, how well these services work, and whether the quality of those services can be further improved. In this work, we conduct a large-scale study to answer these questions. Using RIPE Atlas probes as vantage points, we explore how traffic enters and leaves each cloud's WAN. In addition, we measure the access latency of the WAN-transit and the inet-transit service of each cloud and compare it with that of an emulated performance-based routing strategy. Our study shows that despite the cloud providers' intention to carry customers' traffic on its WAN to the maximum extent possible, for about 12% (Azure) and 13% (Google) of our vantage points, traffic exits the cloud WAN early at cloud edges more than 5000km away from the vantage points' nearest cloud edges. In contrast, more than 84% (AWS), 78% (Azure), and 81% (Google) of vantage points enter a cloud WAN within a 500km radius of their respective locations. Moreover, we find that cloud providers employ different routing strategies to implement the inet-transit service, leading to transit policies that may deviate from their advertised service descriptions. Finally, we find that a performance-based routing strategy can significantly reduce latencies in all three cloud providers for 4% to 85% of vantage point and cloud region pairs.
more »
« less
COMET: Distributed Metadata Service for Multi-cloud Experiments
A majority of today's cloud services are independently operated by individual cloud service providers. In this approach, the locations of cloud resources are strictly constrained by the distribution of cloud service providers' sites. As the popularity and scale of cloud services increase, we believe this traditional paradigm is about to change toward further federated services, a.k.a., multi-cloud, due to the improved performance, reduced cost of compute, storage and network resources, as well as increased user demands. In this paper, we present COMET, a lightweight, distributed storage system for managing metadata on large scale, federated cloud infrastructure providers, end users, and their applications (e.g. HTCondor Cluster or Hadoop Cluster). We showcase use case from NSF's, Chameleon, ExoGENI and JetStream research cloud testbeds to show the effectiveness of COMET design and deployment.
more »
« less
- Award ID(s):
- 1826997
- PAR ID:
- 10158250
- Date Published:
- Journal Name:
- 2019 IEEE 27th International Conference on Network Protocols (ICNP)
- Page Range / eLocation ID:
- 1 to 2
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Modern cloud databases are shifting from converged architectures to storage disaggregation, enabling independent scaling and billing of compute and storage. However, cloud databases still rely on external, converged coordination services (e.g., ZooKeeper) for their control planes. These services are effectively lightweight databases optimized for low-volume metadata. As the control plane scales in the cloud, this approach faces similar limitations as converged databases did before storage disaggregation: scalability bottlenecks, low cost efficiency, and increased operational burden. We propose to disaggregate the cluster coordination to achieve the same benefits that storage disaggregation brought to modern cloud DBMSs. We present Marlin, a cloud-native coordination mechanism that fully embraces storage disaggregation. Marlin eliminates the need for external coordination services by consolidating coordination functionality into the existing cloud-native database it manages. To achieve failover without an external coordination service, Marlin allows cross-node modifications on coordination states. To ensure data consistency, Marlin employs transactions to manage both coordination and application states and introduces MarlinCommit, an optimized commit protocol that ensures strong transactional guarantees even under cross-node modifications. Our evaluations demonstrate that Marlin improves cost efficiency by up to 4.4x and reduces reconfiguration duration by up to 4.9x compared to converged coordination solutions.more » « less
-
The lack of a readily accessible, tightly integrated data fabric connecting high-speed networking, storage, and computing services remains a critical barrier to the democratization of scientific discovery. To address this challenge, we are building National Science Data Fabric (NSDF), a holistic ecosystem to facilitate domain scientists in their daily research. NSDF comprises networking, storage, and computing services, as well as outreach initiatives. In this paper, we present a testbed integrating three services (i.e., networking, storage, and computing). We evaluate their performance. Specifically, we study the networking services and their throughput and latency with a focus on academic cloud providers; the storage services and their performance with a focus on data movement using file system mappers for both academic and commercial clouds; and computing orchestration services focusing on commercial cloud providers. We discuss NSDF's potential to increase scalability and usability as it decreases time-to-discovery across scientific domains.more » « less
-
Transient computing has become popular in public cloud environments for running delay-insensitive batch and data processing applications at low cost. Since transient cloud servers can be revoked at any time by the cloud provider, they are considered unsuitable for running interactive application such as web services. In this paper, we present VM deflation as an alternative mechanism to server preemption for reclaiming resources from transient cloud servers under resource pressure. Using real traces from top-tier cloud providers, we show the feasibility of using VM deflation as a resource reclamation mechanism for interactive applications in public clouds. We show how current hypervisor mechanisms can be used to implement VM deflation and present cluster deflation policies for resource management of transient and on-demand cloud VMs. Experimental evaluation of our deflation system on a Linux cluster shows that microservice-based applications can be deflated by up to 50% with negligible performance overhead. Our cluster-level deflation policies allow overcommitment levels as high as 50%, with less than a 1% decrease in application throughput, and can enable cloud platforms to increase revenue by 30%more » « less
-
The growing demand of industrial, automotive and service robots presents a challenge to the centralized Cloud Robotics model in terms of privacy, security, latency, bandwidth, and reliability. In this paper, we present a ‘Fog Robotics’ approach to deep robot learning that distributes compute, storage and networking resources between the Cloud and the Edge in a federated manner. Deep models are trained on non-private (public) synthetic images in the Cloud; the models are adapted to the private real images of the environment at the Edge within a trusted network and subsequently, deployed as a service for low-latency and secure inference/prediction for other robots in the network. We apply this approach to surface decluttering, where a mobile robot picks and sorts objects from a cluttered floor by learning a deep object recognition and a grasp planning model. Experiments suggest that Fog Robotics can improve performance by sim-to-real domain adaptation in comparison to exclusively using Cloud or Edge resources, while reducing the inference cycle time by 4 to successfully declutter 86% of objects over 213 attempts.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript
- Conference Paper:
- https://doi.org/10.1109/ICNP.2019.8888138
