论文标题
正常时间优化的云体系结构作为经纪服务
Uptime-Optimized Cloud Architecture as a Brokered Service
论文作者
论文摘要
企业工作负载通常要求在滑倒时承担合同罚款的痛苦，要求达成正常运行时间服务水平协议（SLA）。通常，策略是引入临时HA（高可用性）机制。我们进行了调查的实施解决方案不会将其可用性模型映射到所需的正常运行时间SLA和任何预期的罚款。在我们观察到的大多数客户案例中，这要么产生了一个过度设计的解决方案，该解决方案的冗余比所需的更多，要么是在合同中可能会在系统正常运行时间上滑落的解决方案。在本文中，我们提出了一个由模型支持的框架，以自动确定允许的正常运行时间SLA和滑动罚款的TCO（总拥有成本）的支持解决方案。我们试图确定我们的工作最好是推荐正常时间优化的云体系结构的经纪服务。
Enterprise workloads usually call for an uptime service level agreement (SLA) at the pain of contractual penalty in the event of slippage. Often, the strategy is to introduce ad-hoc HA (High Availability) mechanisms in response. Implemented solutions that we surveyed do not mathematically map their availability model to the required uptime SLA and to any expected penalty payout. In most client cases that we observed, this either resulted in an over-engineered solution that had more redundancies than was required, or in an inadequate solution that could potentially slip on the system uptime SLA stipulated in the contract. In this paper, we propose a framework backed by a model, to automatically determine the HA-enabled solution with the least TCO (total cost of ownership) for a given uptime SLA and slippage penalty. We attempt to establish that our work is best implemented as a brokered service that recommends an uptime-optimized cloud architecture.
