Shulou(Shulou.com)06/03 Report--

Before introducing StarlingX, you need to introduce the predecessor of StarlingX. Fenghe established the TitaniumCloud ecosystem project as early as 2014, and the design goal is to become an ideal cloud operating system platform for telecom network applications to run network virtualization functions with the highest reliability. StarlingX is a versatile infrastructure stack for edge computing scenarios. Because StarlingX is generated from open source WindRiver products, and this Titanium Cloud of WindRiver has been developed for 5 years from its establishment to now, Titanium Cloud is a relatively mature product, so even if StarlingX only released the first version in July this year, StarlingX also has a production basis under certain conditions.

StarlingX is both a development project and an integration project. It combines new services with more open source projects into an overall edge cloud software stack. Because Fenghe Titanium Cloud platform has a relatively high improvement in performance and availability. At present, Fenghe's software can achieve carrier-level performance on general server hardware, and the NFV platform can reach 6 9s, which can fully meet the telecom-level requirements. StarlingX is based on the commercial products of Fenghe to optimize the formation of open source, so StaringX also inherits the advantages of Titanium Cloud.

StarlingX is mainly oriented to industrial IoT, telecommunications, video services and other services that require high latency. And based on Titanium Cloud,StarlingX, it provides both performance and high availability.

StarlingX's original code was provided by Intel and Wind River, and the StarlingX project is currently managed by OpenStack Foundation, and StarlingX is currently combined with leading open source projects, including OpenStack, Ceph, and OVS. StarlingX only got its first version of release in October 2018, so StarlingX is a very young project.

For some students who want to learn more about edge computing cloud, the OpenStack community published a white paper on edge computing, introducing the causes, challenges, solutions and application scenarios of edge computing cloud.

The white paper concludes that a mature edge computing cloud is for developers and applications and sinks infrastructure resources onto the edge network. So how do you define edge computing? The most important point is latency, that is, the network latency from all kinds of endpoints to the edge cloud needs to be less than 20ms.

The second part introduces the function point of StarlingX. Because StarlingX contains OpenStack, StarlingX, the function point of OpenStack, also has it, but the following function points are different from OpenStack.

Configuration Management

Configuration management capabilities become very important in the edge cloud infrastructure architecture, especially when managing a large number of remote nodes, because some distant nodes are not easy to configure directly. So with the help of Configuration Management

Function points, you can easily configure the remote physical server management, configuration management includes CPU, GPU, memory, Huge pages,crypto/compression PCIE configuration and so on.

Fault Management

This component can count alarms and view log, and includes both physical and virtual resources of central cloud and edge cloud, and can be viewed on Horizon, monitoring a wider range of aspects than OpenStack.

Host Management

This component can check the status of the virtual host and try to restart automatically when the host shuts down. Different scheduling policies are used to restart the virtual machine according to the cluster status, key processes, resource threshold, physical host failure, and so on.

Service Management

This function point provides high availability of services, uses multiple channels to avoid disconnection of communication and brain fissure of services, ensures high availability of services based on the switching of active/passive status of StarlingX services, and monitors the status of services.

Software Management

Full stack package upgrade from kernel to OpenStack service, which enables rolling upgrades, such as hot migration of virtual machines when physical servers need to be shut down. This function only needs to be operated on the horizon interface in StarlingX. The hot migration can automatically migrate the virtual machines or containers on the hosts that need to be updated to available hosts in advance and after the update is completed. Resources are then automatically allocated to the updated host, which provides a mechanism for life cycle management of virtual machine shutdown during upgrade.

The third part is the overall architecture of StarlingX. StarlingX is based on OpenStack components and enhances its functions, providing the capabilities of the above five core function points.

The full name of OSS is Operation support system, which refers to the operation support system. The full name of BSS is Businesssupport system,BSS system, which includes customer relationship management, data acquisition system, billing and accounting, comprehensive settlement and marketing support. The above function points of StarlingX can empower these systems to provide a more fully functional infrastructure.

The underlying KVM is optimized on the computing node, and the network acceleration schemes of SR-IOV, OVS-DPDK and Intel are introduced in the network part, so that the capability of the computing node is qualitatively provided. if the above function points provide robustness and high availability, the optimization of the underlying components improves the performance of the overall edge cloud. Storage nodes still integrate the industry's excellent distributed storage solution Ceph, and provide a variety of storage solutions to ensure the high availability of carrier-level storage through the integration of distributed, centralized and commercial SAN storage.

Packet core network EPC (EvolvedPacket Core), which is characterized by only packet domain but no circuit domain, based on full IP structure, separation of control and bearer and flattened network structure, which mainly includes MME, SGW, PGW, PCRF and other network elements. Among them, SGW and PGW are often co-located and are called SAE-GW.

CPE, whose full name is Customer Premise Equipment in English, is actually a mobile signal access device that receives mobile signals and transmits them with wireless WIFI signals. It is also a device that converts high-speed 4G or 5G signals into WiFi signals, and can support a large number of mobile terminals that surf the Internet at the same time. CPE can be widely used in rural areas, cities, hospitals, units, factories, residential areas and other wireless network access, can save the cost of laying wired network.

As for the upper virtual network element interface VNFs, StarlingX can provide support for telecom network elements by deploying virtual EPC and CPE in the virtual machine.

The StarlingX community is currently supporting containers and will be able to provide users with both virtual machines and containers at the edge. It will also develop the ability to deploy StarlingX systems in different operating systems.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.