If you need help, please click here:

Open, Hybrid Cloud for Big Data

By Liu Jun, Huawei Senior Cloud Computing Marketing Manager

The future of cloud computing is growing to include public and private hybrids that offer the benefits of both.

Core Assets

Seen as a virtual computer, the cloud is the processor and storage, Big Data is the Database Management System (DBMS), and complex applications are made highly available. Big Data systems collect large volumes of data and extract useful information using data mining and analytics. Big Data applications bring “value” to the cloud.

Data security is a key concern for enterprises to protect their main assets. Core data stored on private clouds has the benefit of improved reliability due to a diversity of approaches, including backup and disaster recovery. Other data is generally carried over the public cloud through consideration of factors such as cost-effectiveness and convenience. However, to better address an enterprise’s business demands, applications and data often need to flow between the two cloud systems. When confronting such issues, CIOs must strive to balance application rollout costs, as well as ensure good compatibility between enterprises’ private clouds and public clouds.

Reducing Enterprise Cloud Costs

CIOs currently face the following challenges: 1) Delivering new services in the quest for differentiated value, and the maintenance of an IT infrastructure flexible enough to support rapid changes in market trends; 2) Balancing performance and cost to accommodate service growth requirements; and 3) Building open platforms compatible with third-party technologies. Platforms must be simultaneously future-proof and protective of existing IT investment.

Backed by years of research i n Internet technologies, the FusionSphere platform is designed to solve these challenges. Administrators can set up a virtual data center within ten minutes using simple visual templates.

The FusionCube implementation of FusionSphere i s the s ingle-package integration of cloud infrastructure delivery for computing, storage, networking, and management. FusionCube offers customers a system at an optimal cost by combining cloud computing and network switching technology, software, and hardware architecture designed for specific applications. On-demand data and application services enable enterprises to achieve more agile and intelligent business solutions.

As a cloud-computing platform, FusionSphere uniformly schedules and maximizes IT resources. With years of technical strength in software and hardware, Huawei’s Research & Development engineers developed FusionSphere to resolve the performance bottlenecks found in traditional IT architectures. FusionSphere was developed to support 80,000 virtual machines on a single management node, and has been ranked first in the SPECvirt virtualization performance test, the FusionSphere management capabilities greatly lower the technical barriers for operating large, cloud-based data centers.

Open Solutions

FusionSphere can integrate the computing, storage, and network resource pools of multiple vendors, and support the hardware of mainstream vendors like IBM, HP, and EMC. In software architecture, it is compatible with the Amazon Web Services (AWS) Application Programming Interface (API) and OpenStack architecture. As a gold member in the OpenStack Foundation, Huawei is continuously contributing high volumes of code to the OpenStack open source platform. Through extensive testing, pre-verification, and pre-integration with mainstream operating systems, middleware, and application software, FusionCube supports mainstream database software to improve system performance without having to modify original applications in the system.

With FusionSphere as the system foundation, Vodafone, a leading Telecom carrier in Europe, selected Huawei to set up a carrier-class hybrid cloud system. Prior to implementing the FusionSphere platform, Vodafone had encountered many technical dilemmas. Initially, to accelerate the development and rollout of applications, Vodafone deployed some of its services on AWS. The results were satisfactory in the beginning, but as the use of services expanded, thousands of applications led to tens of millions of dollars being spent on AWS each year. Consequently, the company’s expenses on the IT platform increased to between 7 and 10 percent of its revenue, much higher than the industry’s optimal level (about 4 to 5 percent).

To reduce its IT platform expenses, Vodafone’s FusionSphere system supports cross-site and cross-data-center applications. Combining OpenStack with AWS API support enables the system to easily migrate applications and data between public and private clouds. Vodafone’s new system helps integrate its 28 global data centers and its AWS public cloud into a single cloud platform. Services that originally deployed on the AWS public cloud can be seamlessly migrated to Vodafone’s private cloud. By doing so, Vodafone’s hybrid cloud solution is now less dependent on AWS with optimal infrastructure costs.

Future-Proof Architecture

Huawei, together with over 400 global partners, jointly explores how best to implement open, advanced, and efficient IT platforms. For example, Huawei and Intel have jointly improved virtualization performance, and provide an industry-leading, multi-vCPU virtual machine backup function. Huawei’s FusionCube team worked with SAP and enabled SAP’s HANA system to accelerate customers’ businesses. Huawei’s desktop cloud team partnered with NVIDIA to achieve GPU hardware virtualization for more complex desktop cloud applications. Huawei also provides an open Software Development Kit (SDK), allowing third parties to develop applications on the FusionSphere platform.

To date, Huawei has delivered FusionSphere solutions to more than 120 cloud computing data centers deployed in 42 countries, and will continue to empower customers to solve the challenges brought by complex applications and Big Data. The future-proof architecture of FusionCube and FusionSphere will be the hybrid cloud platform for enterprises to transparently migrate Big Data as well as add services between public and private clouds. Enterprise data centers are evolving to hybrid cloud solutions, and FusionSphere can fully support flexible deployment of private, public, and hybrid clouds based on customer needs.

Share link to: