Huawei Cloud Fabric Information Center Solution
With the advent of cloud computing, enterprises and carriers of all sizes are focusing on how to build information middle infrastructures capable of supporting sustainable cloud services. Both cloud calculating and data center service providers tin can larn more markets and customers past offering more high-quality, on-demand cloud computing services.
Deject computing helps enterprises lower OPEX and improve the efficiency of O&M. With cloud computing technology, enterprises tin maximize ROI while minimizing resource usage. Co-ordinate to Forrester Enquiry, cloud computing applications are being implemented on an ever-larger scale effectually the earth. Their market size is expected to reach about USD $241 billion by the terminate of 2020. Applications such as Infrastructure every bit a Service (IaaS) and Software as a Service (SaaS) are growing fast and the cloud computing ecosystem continues to mature. For example, Amazon Web Services (AWS) have been deployed in over 190 countries and regions across the globe, with 400,000 or more than commercial users, including The New York Times and NASDAQ.
The rapid development of cloud computing applications is driving the emergence of Software-Defined Networking (SDN). SDN architecture can flexibly schedule information middle resources and provide users with on-demand services. As a result, SDN should accelerate the step of development of cloud computing. Analysts expected the SDN market place to attain USD $two billion by the terminate of 2016.
Increasing deject calculating services and SDN applications are moving information centers towards Big Data, new services, and diversity. However, these new technologies and cloud computing services pose challenges to traditional information centers and their network architectures.
Large Data leads to exponential network traffic growth
Big Data is a fast-growing, massive, diversified information asset. Stronger decision-making and insight as well as higher process optimization are accomplished past using new processing models. Big Data is characterized by huge volume, broad variety, loftier-speed transmission, and significant value.
There are four typical Big Information service architectures: distributed, server cluster, parallel computing, and social media applications. Contempo years have seen unprecedented development of Big Data applications. Gartner, a leading Information technology research and advisory firm, predicts that the average Big Information traffic volume of enterprises around the earth will grow to 800 percent of electric current levels in five years.
Data centers need to scale out to survive the Big Data tide. Network servers will need to grow from 10 GE to forty GE/100 GE. Internal traffic, meanwhile, will keep to grow because interfaces connecting data center networks have evolved to twoscore GE/100 GE. As a event, data centers urgently need a flattened, scalable network architecture that provides non-blocking, high-speed, packet-forwarding capabilities.
Legacy architectures slow cloud data center structure
Enterprise information center management is divided into It and network system direction. Even so, cloud services require not only calculating and storage resource but as well network resources. Less efficient physical network deployment and the separate between concrete networks and IT systems in legacy architectures has become a critical challenge to cloud information heart construction.
More open networking environments required
The original layer 3 network architecture is incompatible with the data heart network architecture. If a bandwidth bottleneck occurs and direction becomes complex, network hardware devices cannot cope with east-westward traffic changes caused by virtualization. To solve this problem, data center networks must be open up and converged. The answer is programmable Software-Defined Networking (SDN).
To help customers quickly adapt to changes in cloud computing services, Huawei offers its CloudFabric Data Heart Solution (CloudFabric). With it, customers are able to build a unproblematic, rubberband, and open deject information centre network that supports sustainable deject service evolution. Huawei CloudFabric uses Huawei’south high-operation CloudEngine (CE) series information center switches based on Huawei’s Versatile Routing Platform version 8 (VRPv8). The solution besides uses Huawei’southward Agile Controller, which uniformly controls and schedules ICT resources and speedily deploys deject services. Huawei CloudFabric is compatible with multiple mainstream cloud platforms in the industry and supports a broad variety of cloud services and applications. This solution applies to Internet companies, financial services organizations, governments, energy companies, large enterprises, and carriers.
Figure i. Solution architecture
Huawei CloudFabric provides multi-layered network products to provision 1-cease-shop services for users and to simplify customer data heart network construction. This solution provides highly scalable, interconnected networks. Taking advantage of its full-series manual, routing, security, and Network Management System (NMS) product families, Huawei provides an all-in-1 data center network solution. In terms of network compages and functions, Huawei CloudFabric consists of the post-obit sub-solutions:
Effigy ii. Solution components
Rubberband scalability provides big pipes for Big Data
Device-level rubberband scalability
Huawei CloudFabric provides information middle network products that feature large switching capacity and continuous, seamless chapters expansion. No device replacement is required. As a result, data center equipment rooms do non demand to be reconstructed. Huawei CE12800 series data eye core switches have the industry’due south highest switching capacity (64 Tbit/s), which can seamlessly calibration to over 100 Tbit/s. A unmarried CE12800 line card has a terabit-level forwarding chapters. The CE12800 provides 12 ten 100 GE super high-operation line cards, providing users with a continuously stable switching core. Huawei USG9500 serial information center security gateways adopt the industry-leading ‘NP + multi-core + distributed’ compages. A single USG9500 bill of fare provides a throughput upwards to 1 Tbit/due south. Huawei USG9500 also provides one.44 billion concurrent connections.
System-level elastic scalability
Huawei CloudFabric seamlessly supports Transparent Interconnection of Lots of Links (TRILL), a standard IETF protocol. A typical TRILL network consisting of core and TOR switches is used to build a large layer 2 network with over 512 nodes. Over eighteen,000 x GE servers can be deployed on a TRILL network. Huawei CloudFabric provides a combined solution of Huawei’s proprietary Cluster Switch System (CSS) and Super Virtual Fabric (SVF). This combined solution virtualizes multiple homogeneous or heterogeneous physical switches into one logical switch, simplifying network management and improving network reliability. CSS is a core switch clustering technology that horizontally (east-west) virtualizes multiple core switches into one core switch. SVF expands heterogeneous core switches vertically (s-north). Multiple leaf switches are virtualized into remote line cards on a spine switch, which flexibly expands interfaces and simplifies cablevision layout and device direction in equipment rooms. Huawei SVF is the only technology that implements local forwarding on foliage switches. Since the majority of the data center traffic is east-west traffic, SVF tin can maximize forwarding efficiency and reduce network latency.
Data heart-level elastic scalability
Huawei CloudFabric provides highly efficient, three-layer connections among information centers, satisfying the interconnection requirements of layer 2, layer 3, and Storage Area Network (SAN) layer services. This solution supports the Ethernet Virtual Network (EVN) Layer two interconnection, implementing cross-region service expansion for upwardly to 32 data centers beyond IP Wide Expanse Networks (WANs). EVN integrates resources of multiple information centers into a super large Information technology resources puddle. This integration not only supports server clusters across multiple data centers, just as well enables up to 256,000 Virtual Machines (VMs) to smoothly migrate across data centers. In addition, by leveraging Huawei’due south 400 GE core routers and 100 GE Wavelength Division Multiplexing (WDM) products, Huawei CloudFabric provides a highly reliable ‘IP + Optical’ solution for disaster recovery, ensuring not-stop data centre service traffic forwarding.
Open up ecosystem
Huawei has been proactive in building a deject calculating data center ecosystem. Huawei cooperates with world-renowned IT vendors and standards organizations, enabling the Huawei CloudFabric to seamlessly connect to mainstream cloud platforms such as VMware vCenter, Microsoft System Heart, OpenStack, and Huawei FusionSphere. In this fashion, network and IT resource can be managed in a unified manner. For example, Huawei teamed up with Microsoft to launch a hybrid network virtualization solution. This solution uses Huawei CE12800 as a hardware virtualization gateway to provide high-performance communication between Microsoft VMs and campus and co-operative users, too as betwixt Microsoft VMs and VMs and physical servers from other vendors. Additionally, Huawei is Red china’southward start gold member of the OpenStack Foundation and is committed to promoting SDN industry concatenation evolution.
The Agile Controller can parse the service language and connect to multiple mainstream cloud platforms to provide a deject application-oriented network. Users can uniformly schedule network, computing, and storage resources on demand on the cloud platforms.
Because information technology is service-axial, Huawei CloudFabric permits IT administrators to hands schedule network resources and implement cloud-based network migration. IT administrators can utilise the service language to ascertain and adjust network requirements (application profile views) from the service perspective. Each type of service has an contained awarding contour view. The Agile Controller tin can understand three types of views, automatically convert the application profile views into logical networks, and evangelize configurations to physical networks. Every bit a effect, network resource can be migrated on need.
The Agile Controller integrates multi-tenant network management, automatic orchestration of service chains, and VM access control. It centralizes control and dynamic scheduling of ICT resources, as well as fast cloud service deployment, and features superior openness. Its northbound interfaces tin connect to a variety of mainstream deject platforms from other vendors, ensuring that a cloud platform can hands deliver and adjust network policies. Taking reward of open APIs and OpenFlow standards, the Active Controller southbound interfaces can control and manage both Huawei and third-party network devices. The open Agile Controller not only improves users’ service commitment efficiency, but also facilitates network customization based on specific service characteristics, implementing seamlessly converged networks and services.
Huawei CloudFabric besides provides device-level openness, offer diversified choices to customers. Huawei CE series data eye switches support open APIs and OpenFlow standards, directly connect to mainstream platforms and 3rd-party controllers, and build open, controllable networks. Running on a fully programmable network using Huawei’s Ethernet Network Processor (ENP) chips, Huawei CloudFabric helps customers apace develop and deploy new services.
ten times faster cloud service provisioning through cloud-based network migration
Cloud service connections:
service-driven solution enables it administrators to easily schedule network resources, implementing cloud-based network migration
The Huawei CloudFabric compages helps customers build rubberband, simple, and open up deject information centers. This architecture also provides agile, innovative solutions that enable networks to quickly adapt to rapidly changing cloud services, and supports connected cloud service development. Huawei CloudFabric also is capable of fast development of enterprise cloud services over the side by side ten years or longer. Other benefits include:
Figure iii. Solution benefits
Elastic: Flexible Cloud Service Scalability, Providing Big Pipes for Large Data
Huawei Cloud Textile provides device-level, system-level, and data centre-level elastic scalability for data center networks, and is capable of satisfying high-speed development requirements of enterprise cloud services over the next x years or longer.
Figure 4. Elastic cloth
Figure five. Application framework and logical and physical networks deployed as It resources
IT and network administrators tin can assess global network resource distribution, resource usage, and fault condition in real time.
Figure 6 Separated vs. mutually-enlightened concrete and virtual networks
The Agile Controller can accurately locate network failure points in a timely way and adjust physical and virtual networks every bit needed. It tin also associate applications with the logical network. In this manner, upper-layer applications are aware of network faults, ensuring quick service optimization.
IT administrators tin focus on logical networks and flexibly customize logical network policies based on service requirements. Logical networks shield the differences between various physical and virtual networks and networking technologies. As a consequence, Huawei CloudFabric is compatible with any physical network model, avoids reinvestment caused past the need to repace legacy devices, thereby reducing TCO.
Figure seven. Decoupling logical networks from concrete networks enables It administrators to customize services
Effigy 8. Huawei Cloud Fabric seamlessly connects to mainstream deject platforms
Figure 9. Open fabric
Huawei has 20 years’ experience in the IP field and an industry-leading series of network products and solutions. Information technology as well is recognized as a leading global network solutions provider. Huawei’southward world-class inquiry capabilities and experts offering unparalleled experience in the areas of network standards and flake development. As a fellow member of several international standards organizations, for example, ONF, IETF, and IEEE, Huawei contributes to standards inquiry, product development, and customization capability improvements. Huawei too remains committed to providing intelligent, programmable, and open networks to customers through its carrier-grade network experience and innovative products.
Equally illustrated in the following figure and table, Huawei Cloud Fabric provides network products at unlike layers.
|Data Center Switches||CE12800 series||Manufacture-leading data centre switches|
|CE7800 series||Access or assemblage switches that provide full 40 GE ports|
|CE6800 series||Box switches that provide loftier-density 10 GE ports to connect to downstream devices and twoscore GE ports to connect to upstream devices|
|CE5800 series||Box switches that provide high-density GE ports to connect to downstream devices and twoscore GE or 10 GE ports to connect to upstream devices|
|Security Products for Data Centers||USG9500 series||High-performance firewalls|
|USG6000 serial||Next-Generation Firewalls (NGFWs)|
|AntiDDoS series||DDoS attack defense force devices|
|Routers||NE40E serial||Egress routers for data centers|
|Optical Transport Network (OTN) Products||OSN serial||OTN devices for information centers|
|Intelligent NMS for Data Centers||eSight||Visualized intelligent network direction platform|
|Data Center Network Controller||Active Controller||Data center network controller|