Data can enable the insights that lead to better decisions and organizations to be more effective. But reaching these insights often requires a high performance network and integrated infrastructure that’s ready for analytics workflows. Organizations that use their data and turn it into strategic information will lead their organizations through the current business intelligence transformation and help secure their future.
For example, one large East Coast State University conducts $150 million of federally funded research annually—all of which relies on the university’s network. Big Data and analytics are critical to their research and mission. University departments and colleges use centralized IT support to obtain networking infrastructure and services, such as voice over IP (VoIP), wireless, administrative computing, research computing, and Internet connections. Today, that network includes 4,000 LAN switches, 5,000 wireless access points, and 18,000 VoIP phones – and it is growing rapidly.
The university’s research network is a work in progress and as the university finishes its research and science network migrations, it will create a high-performance network optimized for research traffic and research science flows. Already, large research data transfers that used to take up to 15 days have been reduced to as little as 20 hours. And with superior visibility into network traffic, sources, and destinations, the network team is reaching out to research users of the network and helping them accelerate results, gain more time for experimentation, and save money.
University network changes have achieved some unique results:
Optimized a high-performance, world-class research network
Converged infrastructure to deliver network capabilities with superior user experience across campus, state, and international locations
Simplified management of network resources, saving time and operating budget
Greater visibility into network traffic for optimizing scalability, planning, decision making, and internal customer service
Unique requirements for compliance and new data lake analytics applications like Hadoop, demand ever-increasing volumes of unstructured data to be collected, archived and available. As such, their network requires scale-out capacity while effectively managing capital and operational expense. The features in Brocade VDX switches that align to these performance and scale needs are:
Deep on-chip buffers
Consistent network bandwidth and low latency
Highly scalable Leaf/Spine non-blocking topology
Automated provisioning, troubleshooting and management
In addition to these, Brocade VCS Ethernet fabrics have a distributed control plane which minimizes flooding in the network, enabling better overall application performance. These performance features directly and positively affect the performance of Data Lake analytic solutions.
Large IT organizations typically need to scale to address large analytic workloads very quickly in order to address the needs of their growing unstructured Data Lakes. Just like Dell EMC Isilon’s ability to simply scale the Isilon cluster, Brocade Ethernet fabrics also easily scale.
Scalability: Brocade VDX Switches and VCS fabric technology is designed for simple network scalability. Just plug a new switch into the existing fabric and it is automatically configured. This automation enables scaling network capacity as requirements demand. Even adding additional bandwidth between switches requires no configuration at all. You simply connect new cables, frame-based Inter-Switch Links (ISL) trunks form and traffic is automatically redistributed among the trunk members, with no impact to applications. In addition, Brocade VCS fabrics employ unique multipathing capabilities in both hardware and software to efficiently use the network for best performance and maximum network utilization, allowing organizations to elastically grow domains with optimal network utilization through efficient, dynamic load balancing at Layers 1–3 for maximum network resilience, bandwidth, and scale.
Manageability: Logical Chassis functionality in Brocade VCS fabrics extends the simplicity of VCS fabric operations by streamlining fabric management and maintenance. This management allows the use of a single “virtual IP address” to perform all configuration fabric-wide. The same IP address is used to upload new software and to auto-provision the cluster. It enables central management from a switch of your choice, while allowing for granular visibility and control. Eliminating repetitive, manual commands reduces opportunities for error and downtime, and the system-wide view facilitates fast, easy troubleshooting. It saves administration time and money by simplifying network management. And it provides a degree of management abstraction that simultaneously simplifies engagement with higher-level orchestration tools while remaining intimately aware of every aspect of the fabric. For example, In VMware environments, there is tight integration with vRealize operations and the VCS fabric at the network layer, sharing event and status information, syslog and VLAN information with vRealize to assist VMware administrators in root cause analysis of VMware security and performance issues.
Brocade VDX Switches offer an advanced feature set that Big Data environments require. With 10/40/100 GbE options for designing oversubscribed or non-oversubscribed networks, high throughput, and optimized buffer and latency, the Brocade VDX is an ideal switch for Big Data applications. Together with Brocade VCS Fabrics, Brocade VDX Switches can simplify network design and operations for both cloud and Big Data network fabrics.
As enterprise big data workloads increase in size and complexity, the network will play a more and more crucial role in ensuring workloads are completed and insights are timely delivered. Brocade seamlessly integrated with EMC storage offers an ideal solution to address network challenges caused by increasing network traffic from Big Data and analytics.