文件名称：Windows Azure Succinctly(Syncfusion,2014)
文件大小：2.91MB
文件格式：PDF
更新时间：2018-04-09 03:37:49
windows azure cloud
most commonly driven by the management tasks distributed among the host platform and customer. These categories (described in more detail later) include Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). For the purposes of discussion, we’ll start with the assumption that a typical enterprise application has the following set of concerns at the infrastructure level (shown in Figure 1): Network: This includes the physical equipment such as switches and routers as well as the logical configuration of subnets. Many application developers often take the network part of the solution for granted, but there is no amount of code that can overcome the devastating effect of a poorly configured network on the stability and performance of an application. Storage: Most software applications and the businesses they support are highly dependent on the storage of data and in keeping that data safe from loss and unauthorized disclosure. Management tasks at the storage layer include configuring storage devices for optimum throughput, monitoring for device failure, tracking utilization, and ensuring that a backup plan exists and is followed. Servers: While my earliest memory of managing a network included wall-to-wall tower computers, each with a special function, the modern data center is characterized by fewer and larger physical machines, each hosting a number of virtualized servers. This model is much easier to manage than the model without virtualization, but it still brings a number of management tasks such as monitoring the servers for hardware failure, ensuring that adequate power is available for the equipment, and maintaining environmental conditions for optimal performance of the hardware. Virtualization: As noted in the discussion of physical servers, virtualization is the norm in the modern data center. Products such as Microsoft’s Hyper-V, VMware’s vSphere, and Citrix’s ZenServer all let you perform management tasks such as managing the 11 distribution of physical resources among the virtual hosts, monitoring to optimize utilization of hardware, and provisioning and decommissioning of virtual hosts. Host Operating System: Virtualization does not remove the management tasks of host operating systems. These still need to be appropriately configured for performance, availability, and security. Management tasks often include ongoing application of operating system and security software (such as antivirus and intrusion detection) patches and any operating system-specific monitoring tasks that cannot be performed by the virtualization tool. Middleware: Middleware includes software that is used as a foundation on which to build applications without having to reinvent functionality commonly found within specific applications—allowing developers to focus on the specific value intended to be provided by their application. Common examples of middleware used by applications built on the Microsoft stack include (but certainly is not limited to) Microsoft Internet Information Services (IIS) (formerly Internet Information Server) to handle tasks such as managing request/response for HTTP requests, Microsoft AppFabric for providing a distributed cache, and Microsoft SQL Server to handle storage and retrieval of relational data. Each middleware comes with its own set of configuration and management tasks that varies by the complexity of the particular middleware, ranging from the “set it and forget it” nature of Microsoft AppFabric to the constant monitoring and tuning provided by database administrators for some Microsoft SQL Server implementations. Application Run Time: The application run time includes the container in which applications execute. For web applications built on the Microsoft stack, this container is most commonly Microsoft ASP.NET running either ASP.NET Web Forms or ASP.NET MVC. Certain run time configuration parameters (such as processing and memory resources to allocate to the container) are managed at the host level while other parameters are available for customization by the individual application. Application Data: The application data is usually the tangible product of having executed the application. Whether it’s recording bills of sale for a multinational product supplier or high scores in the latest bubble app craze, the management of what data to store and how to keep it consistent with the applicable rules is usually performed by the application code itself. Therefore, very little management activity that is not performed either at the storage or middleware layers is necessary. Application Code: The application code is the executable code that executes within the run time to perform actions directly related to achieving the goal of the application. This is typically the layer of the application stack most visible to the business stakeholders of the application. It is the layer on which the code most specific to solving problems related to the business domain can be found. I am often heard telling co-workers and clients that developers should work on solving problems that only they can solve. By that I mean developers should focus on the code in the application that is most focused on solving the specific business problem, instead of trying to reinvent frameworks for solving problems that are already addressed in common frameworks or by the run time.