Introduction and application of IBM SmartCloud provisioning products

For such a new product that covers the key technologies in cloud computing, the most popular word in the It circle, this article focuses on describing its features and architecture, how to deploy it, what features it provides to users, how users can use it effectively to reach their needs, and so on.

The main features of Ibm®smartcloud provisioning (ISCP) are high scale and low touch. The so-called high scale refers to smartcloud provisioning with very rapid deployment capabilities. It employs end-to-end peer-to-peer parallel and distributed technology that allows multitasking to be performed in parallel to enable the ability to manage thousands of of servers, allowing one build to be completed in minutes and one hours to build thousands of virtual machines; low touch refers to SmartCloud provisioning With very convenient operation management, so that the objective requirements of management and control to minimize, provide automatic recovery and high fault tolerance so as to greatly enhance the convenience of user management.

Below is the SmartCloud provisioning infrastructure diagram.

Infrastructure

Diagram 1.ibm®smartcloud provisioning topological structure diagram

As can be seen from the topology map, a typical Smart Cloud provisioning system environment consists of two types of servers, compute nodes (Compute node), and storage nodes (Storage node). Storage nodes focus on disk and network I/O, providing image and file services for managing all storage-oriented resources and tasks; compute nodes provide CPU and memory resources for hosting deployment virtual machines, handling all computing-oriented resources and tasks. Other servers include: PXE server (also known as First-box) provides DHCP, DNS, HTTP, and TFTP services to quickly build SmartCloud provisioning components on all networks; Apache Zookeeper A child of Hadoop that provides a distributed application coordination service; HBase is a child of Apache Hadoop, a distributed, column-oriented open source database that provides distributed storage systems; Web Service through the RESTful Web Services Interface, Receiving requests from managers and users, handing the work over to the cloud; the console (Web console and Web Admin Console) is provided to end users and administrators to operate the user interface; LDAP Server is responsible for user authentication management; Virtual Image The Library provides image management services; ICON (image Construction and Composition Tool) helps users build the underlying image, which typically contains the operating system and other software. Typically, compute nodes and storage nodes need to be deployed on physical machines, PXE Server is installed on physical machines, and zookeeper, Hadoop + HBase, Web Service, and Web Console and Web Admin console can be installed on virtual machines , running on a storage node or on a (host) physical machine on a compute node.

Deployment implementation

In a typical minimized deployment architecture, four machines need to be prepared (at least one of the four machines that supports virtualization technology is a compute node), and these four machines are connected to a LAN through a high-speed switch. The connection device recommends a high speed switch to ensure rapid real-time interaction between the nodes of each server. A typical minimized deployment architecture is:

Figure 2. A typical minimized deployment architecture

In the smallest architecture shown in Figure 2, the PXE server, compute node compute, storage node storage-1, and storage-2 are all installed on the physical machine, while the other services are distributed as virtual machines in compute node compute, storage node storage-1, and Storage-2 includes 3 Zookeeper services, 3 HBase services, 1 admincosole services, 1 webconsole services, 1 WebService services and one Ldap-server service. The roles of each node and service are described in detail in the previous paragraph and are not discussed here.