IT is playing an increasingly vital role in ensuring enterprise remain competitive in the cloud computing and Big Data era. To enhance efficiencies, enterprises are gradually moving their core services to the cloud. However, migrating databases and core applications poses new challenges to virtualization technologies, system performance, data reliability, and manageability. At the same time, the migration demands more system and application management personnel and higher skill sets. These new challenges are driving evolutions from conventional IT infrastructures to hyper-converged layouts.
Rapid Development of Hyper-Converged Infrastructures Turbocharges Migration to the Cloud
Hyper-converged infrastructures provide an ideal solution to solving many of the difficulties in the migration to the cloud. The hyper-converged infrastructure market has grown much faster than the industry expected, stealing the spotlight in enterprise-class IT product offerings. Converged infrastructures are now evolving into the second generation – hyper-converged infrastructures. Currently, most enterprises still use conventional IT system solutions in which servers and storage assets are separately deployed. In the first-gen deployments, system integration was the main focus of the convergence to solve the problems in managing dispersed IT devices in the data center. Admittedly, first generation convergence in infrastructure, Convergence 1.0 as it is affectingly called, solved some deployment and management complexities but left plenty of room for improvement in efficiency and overall performance. Enter Convergence 2.0 – the push towards hyper-convergence. These new systems deliver exciting innovations in enterprise IT builds, providing the proverbial key to the new superhighway.
Simply growing performance markers alone cannot accomplish the evolution from convergence to hyper-convergence. Hyper-converged systems feature optimizations to the software layer and deeper levels of integration at the hardware level, even transforming the infrastructure into a unified resource pool that provides computing and storage capabilities. The ‘infrastructure in a box’ model changes how resources are controlled, managed, and even utilized at the basic level. Enterprises can easily deploy cloud and Big Data applications to hyper-converged systems. The added level of integration and maneuverability achieves convergence in software and hardware. The optimization that breaks through traditional thinking in computing and storage architecture layout, achieving enhancements in performance and efficiency, reducing costs, and simplifying O&M. Put simply, hyper-converged infrastructure moves into an organic framework from the original model of simply stacking more devices. The new synergy provides enterprises with the thrust to forge ahead on more ambitious cloud roadmaps.
Four Defining Attributes: Agility, Simplicity, Responsiveness, Openness
Attaining to the needed levels of agility, simplicity, responsiveness, and openness in hyper-converged builds for enterprises is what will really fulfill the true requirements of business operations. Targeting improvements in database and virtualization applications, hyper-converged approaches take computing and storage resources and combine them into one. Pre-integration of distributed storage engines, virtualization platforms, and cloud management software allows enterprises to easily deploy the most agile database platforms and virtualization environments.
‘Agility’ refers to flexible deployments. The size of the equipment room for most enterprises means only a limited number of server, storage, network, and other modules can be deployed, and the percentages of which appliances will be required is often rather unpredictable. Some enterprises may require more capacity while other may experience sharply escalating requirements on performance. This situation creates pain points in managing the sprawl, not to mention planning for upgrades, carrying out device replacements, and other time and resource sinks. Therefore, hyper-converged infrastructures require high integration among computing, storage, virtualization, management, and multiple other functions while fully considering the needed supports for whatever upgrades may be needed in the future.
‘Simplicity’ points to the ease of system installation, deployment, and expansion. Enterprises have a limited number of IT personnel. In even the more basic IT architectures of the past, configuring virtual machines and the ensuring complexities in keeping the system afloat in day-to-day operations kept IT departments plenty busy. When an upgrade was needed to gear into business expansions, those resources became stretched even farther under the tremendous workload of managing the current IT while incorporating or switching to new devices and systems. Hyper-converged infrastructures enable the fastest installations, deployments, and expansions all with reduced learning curves and skill set requirements, lessening the burden on enterprise IT departments.
‘Responsiveness’ mainly speaks to performance enhancement requirements in database applications. Enterprises tend to make extensive use of databases in their operations. Naturally then, these performance-intensive applications are also among the most common. Such high performance requirements means hyper-converged architectures must deliver powerful performance, which does not just mean faster computing speeds because superior counters in storage system response must also be demonstrated. Most database applications require high-performance disks for read/write operations. The I/O speed of the storage system is the bottleneck for most of these types of applications. Achieving hyper-convergence means the storage system must be able to pool resources at the software definition level and deliver higher IO throughput for the database applications to ensure overall performance.
‘Openness’ is a key consideration for enterprises when deploying their IT systems. The market is full of virtualization platforms, but most of current offerings force IT decision-makers to select one specific to their business or service particulars, leaving little flexibility should the dynamics change. Any solution claiming the hyper-convergence credential should feature an open infrastructure if it truly intends to provide a wide-reaching response to meet the current service dynamics, gear into the existing IT environment, and accommodate future expansions.
Rapid business developments require infusions from IT enablements. The four attributes of hyper-converged infrastructures (agility, simplicity, response, openness) help enterprises edge out their advantages in the intensifying market competition. Hyper-converged infrastructures empower enterprises to focus on their established businesses and expand into new areas and customer bases.
Huawei FusionCube: Enabler for the Move to the Cloud
As a supporter in open architectures, Huawei showcased its latest updates to the enabler of the move to hyper-converged infrastructures for enterprises at CeBIT 2016. Named Huawei FusionCube, the solution brings entirely new possibilities in deploying IT assets. The in-house innovation added to the impressive list of recent high-impact launches boasts the most powerful engine.
Based on the same hardware platform, FusionCube 9000 and FusionCube 6000 are developed to cater to different workloads in database and warehouse acceleration, hybrid workloads, and virtualization scenarios. FusionCube 6000 is geared for smaller customers that need to build and manage cloud and virtualization environments with the ability to quickly scale as their news grow. FusionCube 9000 is for larger customers that need high performance and scalability.
The ‘Cube’, as it is sometimes called, adopts three key elements to accelerate database and data warehouse applications and optimize overall performance. First, use of FusionStorage distributed storage systems eliminate engine bottlenecks, improve storage bandwidth, and enhance random read/write capabilities. Using InfiniBand switching achieves up to 56 Gbps on non-blocking networks while PCIe SSD cards accelerate data read/write capabilities. Along with the breakthroughs in overall performance, the cost-effectiveness of the overall database solution is also greatly improved. In an actual use case concerning Sinopec in which an the Huawei FusionCube SAP HANA solution was deployed, the highly effective IT platform improved database load performance by 60% and shortened data uploads times from four hours to thirty minutes. The processing speed also increased many times over, marking four- to six-fold improvements in many metrics. These advances lead to substantial gains in sale reporting, financial analysis reporting, decision-making, enterprise operations, and many other aspects of the business.
In terms of the virtualization applications, Huawei FusionCube offers an enterprise private cloud architecture, essentially forming the basis of the new ‘IT in a box’ model. Proprietary management software achieves unified management and simplified O&M. The pre-integrated, validated, workload-optimized solution greatly shortens the deployment cycle. The scale-out architecture allows for modular builds enabling rapid capacity expansions, much like working with wooden building blocks. Enterprises can now deploy cloud services with the agility and effectiveness beyond what was previously possible. As a tremendous advocate of an open ecosystem, Huawei’s hyper-converged systems support a wide range of platforms, supporting such self-developed systems as FusionSphere in addition to virtualization platforms from other vendors, such as VMware and KVM. In another case of demonstrated success, Huawei FusionCube helped COFCO Coca Cola build their private cloud platform. The build pooled resources into one platform to enhance the full spectrum of core service systems, including everything from ERP and CRM to OA, enterprise mailbox, and HR subsystems. Operational efficiency improved by 300% and TCO dropped 40%, and the solution occupied 70% less space in the equipment room. The industry-tested FusionCube delivers on its promise to make best use of IT resource while enabling flexibility in how overarching system assets are used and which systems can use them.
The Huawei FusionCube has been widely applied in the public sector, finance, transportation, manufacturing, education, retail sales, energy, telecom, and a long list of other industries. It creates exceeding value for customers, a true enabler in the moving of services to the cloud for enterprises.
Download the White Papers, Case Studies and Free Trial of FusionSphere Foundation to learn more how Huawei Cloud Computing Platform helps Enterprise in Digital Transformation.FusionSphere Foundation is a free version of Huawei FusionSphere, including FusionCompute and FusionManager, which provide server virtualization and cloud management capabilities. For a system with less than or equal to six physical CPUs, the Foundation Edition allows users to use FusionSphere free of charge permanently without any function restrictions.For a system with more than six physical CPUs, this edition allows users to use FusionSphere free of charge for up to 90 days without any function restrictions. After the trial period (90 days) expires, existing VMs in the system are still available. However, administrators can only query service data, and manage users, roles, and license files.