Emulex Blog: Emulex Labs

A ‘Great Debate’ on NVGRE vs. VXLAN at Interop

Posted May 16th, 2012 by Emulex Labs Team

Author: Brandon Hoff

Today, at Interop, Emulex and Arista Networks went head to head on a panel discussion covering Network Virtualization using Generic Routing Encapsulation (NVGRE) vs. Virtual eXtensible Local Area Network (VXLAN). We held this debate in front of about 200 people at Interop in the Data Center Networking Track on Monday.

The debate format required each vendor to take an extreme position, e.g. NVGRE is the best or VXLAN is the best. Here were some interesting details:

How many people have heard about VXLAN or NVGRE? Here is my estimate of the responses.

For this group, it is important to note that about 85% of the respondents haven’t heard of either NVGRE or VXLAN (or didn’t respond). I have also seen this at the end user level as well. Out of the last 50 end users I have talked to, most had said that they hadn’t heard of these new protocols, but, they are hungry to learn more.

As we discussed in the debate, the key requirement for either standard to be successful is the ecosystem of products that support it. Given that open vSwitch, VMware, Cisco, Emulex, Arista are supporting VXLANs and Microsoft, Arista, Emulex are supporting NVGRE, VXLAN has the edge with VMware and Cisco supporting it. When asked which standard will win, the vote was 30% VXLAN, one person for NVGRE, but nearly 70% didn’t declare a winner. This is probably the most important takeaway from the debate – there is no clear winner today. It is still early in the market for Overlay Networking, but, it is great that the vendors and the community here at Interop 2012 are taking time help educate people.

Note, none of this is scientific, but this is a good indicator that even though Overlay Networking solutions are being released this year by the major hypervisor vendors, there is some work to be done to educate end users on what Overlay Networks are, how to manage, architect and deploy Overlay Networks.

Introducing New I/O Solutions for the High Performance Computing Market

Posted May 8th, 2012 by Emulex Labs Team

Today at Interop, we unveiled a new family of High Performance Networking (HPN) products that addresses the growing High Performance Computing (HPC) market, key vertical market segments, and the changing needs of today’s data center.

Emulex announced its new family of OneConnect OCe12000 10Gb Ethernet (10GbE) Network Xceleration solutions, which are being brought to market in partnership with Myricom, an industry leader in delivering extreme performance Ethernet solutions for vertical market applications. This multi-generational strategic partnership leverages Myricom’s deep expertise in HPN, specifically low-latency, high packet-rate, vertical application optimizations, and Emulex’s broad enterprise reach, high volume, enterprise-hardened hardware roadmap, and global go-to-market channels.

The new Emulex OneConnect Network Xceleration solutions provide extreme performance for specific enterprise applications that rely on low latency, superior message rate or packet pacing through FastStack™ application acceleration software. Using FastStack acceleration, customers get the benefit of extreme networking performance for their applications, within industry-standard 10GbE. Going forward, Myricom and Emulex will be jointly developing additional FastStack software solutions which will optimize Emulex’s I/O connectivity solutions across additional vertical markets.

These application-specific solutions, including FastStack DBL™, FastStack Sniffer10G™and FastStack VideoPump™ target vertical markets that require extreme performance to meet customer, business or regulatory requirements: High Frequency Trading (HFT), network security, and cloud-enabled video distribution. Not only are they application-specific, they are extremely fast. In fact, they’re so fast that we’re demonstrating their performance capabilities at Interop this week, in our booth #1117. Check out my colleague Mark Jones’ blog over in the Implementer’s Lab for more details on our demonstrations at the show this week.

And if you’re at Interop, you’re in for an extra treat. Not only are we demonstrating extreme performance, we’re giving it away, too. Come by our booth to find out how you can win a Ducati Monster 696. Talk about performance!

OneCommand® Vision + Microsoft® System Center 2012 = Total Data Center Management

Posted April 19th, 2012 by Erick Crowell

Microsoft (MS) System Center 2012 transforms the way you deliver IT and application services to your organization, optimizing resources across public and private cloud, managed from a single pane of glass. This week, we announced support for MS System Center 2012 with Emulex OneCommand Vision, enabling total data center management. OneCommand Vision and MS System Center 2012 target increasing server and application performance and availability while decreasing operational expenses (OPEX). Delivering time-saving features, OneCommand Vision extends its unique application view of I/O performance and availability to MS System Center 2012 reporting critical performance violations, based on historical performance data and intelligent alert analytics. This includes OneCommand Vision’s industry leading service level agreements (SLAs) that help enforce the service, performance and availability of critical storage resources.

With OneCommand Vision and MS System Center 2012 administrators can quickly determine if storage I/O performance is the source of application performance degradation, often simply ruling out storage so administrators can focus their investigations on the server or application. Proactively addressing I/O issues before they result in application downtime is paramount to reducing OPEX just another major benefit provided by this combination.

So how does this combination work? As shown in the figure below, OneCommand Vision forwards SNMP Traps (aka “alerts”) to one or more MS System Center 2012 systems when pre-configured Intelligent Alerts in OneCommand Vision are triggered. The OneCommand Vision web application is used to configure the specific performance and health metrics used in triggering alerts. In fact, this time-saving integration takes just a few minutes for administrators to quickly associate groups of storage resources with one or more I/O performance policies and thresholds. MS System Center 2012 can then log the alerts or integrate them into action scripts that can further automate the troubleshooting process, saving you precious time in resolving or preventing I/O performance or availability issues.

This integration helps to lower the average time to resolution (TTR) of tricky application performance issues thereby lowering, or even preventing, unplanned OPEX. Used across Windows environments, MS System Center 2012 delivers end-to-end service management, working seamlessly with Microsoft infrastructure servers, such as Windows Server, and application servers, such as Microsoft Exchange and SQL Server, helping to increase efficiency while enabling greater control of the IT environment.

The combination of Emulex OneCommand Vision and MS System Center 2012 provides you a single point to manage your infrastructure with robust I/O management capabilities, reducing your OPEX for use in other critical areas of the data center. Try MS System Center 2012 with OneCommand Vision for a time-saving experience today that lowers OPEX through prevented outages tomorrow.

Emulex Rocks! 8GFC dual- and quad-channel HBA leads performance

Posted April 13th, 2012 by Barbara Porter

We’ve just released a new whitepaper detailing the high-performance of Emulex LightPulse 8Gb Fibre Channel (8GFC) dual- and quad-channel adapters, and the results? Blazing-fast! With so many data centers running out of ways to perk up their performance, Emulex LightPulse 8GFC adapters offer a sure bet. With Emulex 8GFC Host Bus Adapters (HBAs), you can get 400,000 input/output operations per second (IOPS) on the dual-channel HBA and over 650,000 IOPS on the quad-channel HBA!

Applications, such as online transaction processing, database applications and server virtualization, have the highest need for data throughput. Dual-channel and quad-channel performance is important as a dominant configuration, with many blade servers standardizing on multi-channel HBAs which also meet virtualization requirements (ports/virtual machine ratios) and deliver high availability.

Emulex LightPulse 8GFC dual-channel and quad-channel HBAs significantly outperform the nearest competitor, as shown in our latest performance whitepaper based on Emulex Labs testing. Some of the findings include the following:

  1. Emulex LPe12002 dual-channel 8FC HBAs offer the highest performance for critical business applications over the QLogic QLE2562:
    • 127% better second channel scalability as measured in IOPS
    • 37% greater total IOPS when saturating both channels
    • 33% more effective CPU utilization
    • 15% greater transaction performance
    • 115MB/s more throughput
  2. Emulex LPe12004 quad-channel 8GFC HBAs offer the highest performance for critical business applications over the QLogic QLE2564:
    • 58% greater total IOPS when using four channels
    • 20% more effective CPU utilization
    • 33% more throughput on read/write operations

It’s no wonder Emulex scored a 96% approval rating for product performance in 2011 (Emulex Labs customer survey) and was selected to power the #1 fastest database system in TPC-C testing (TPC-C Benchmark reports, December 2010). Download our whitepaper to discover what Emulex LightPulse 8GFC dual- and quad-channel HBAs can do for you.

New Deployment Guide: Ideal for Users Deploying VMware EXSi 5.0 with HP 8Gb Fibre Channel.

Posted April 5th, 2012 by Emulex Labs Team

As many of you know, VMware ESXi 5.0 adds a broad range of features to the hypervisor; however, taking full advantage of these features requires the higher speed of an 8Gb SAN in your virtualization environment. To help you transition to an infrastructure that can support the storage needed by the new generation of virtual machines (VMs), Emulex recently published a Solutions Implementer’s Series deployment guide, validating the functionality and performance of ESXi 5.0 in conjunction with 8Gb Fibre Channel (8GFC) connectivity. The tested environment included the best-of-breed HP ProLiant DL 380 G7 server equipped with a dual channel 8GFC adapter produced for HP by Emulex. Storage was provided by an HP 3PAR Storage array.

In addition to recommending that the higher bandwidth 8Gb SANs are used for the new release of vSphere 5.0, the deployment guide outlines the process for deploying ESXi 5.0 with 8GFC connectivity and highlights the results of performance tests carried out in such an environment. First time users of VMware with HP’s 8GFC adapters will find this solution provides ease-of-use, simplicity and supportability. VMware administrators using OneCommand Manager for VMware vCenter have a single glass pane view for centralized management of the adapters. This allows data centers to fully leverage savings in equipment, energy, data center footprint and administration costs that result from server virtualization.

Upgrading to 8GFC will increase your storage bandwidth and performance while maintaining backward compatibility with your existing 2GFC and 4GFC technologies. In addition, you will see the benefits of centralized and streamlined adapter management. Testing carried out by Emulex indicated that an 8GFC implementation based on the HP 82E 8GFC HBA was able to support more VMs than a 4Gb implementation, while also reducing CPU utilization on the ESXi 5.0 host.

Planning the deployment of an ESXi 5.0 host – or the migration of an existing host to ESXi 5.0 – is a simple process. Using 8Gb storage, the resulting infrastructure will be able to meet the demands of the larger VMs you can now create and will allow you to introduce additional business-critical applications.

Bottom line: If you are a first time user, or contemplating the move, this deployment guide is for you! It doesn’t get any easier!

The Implementer’s Lab Solutions Implementer’s Series deployment guide can be downloaded here:
VMware vSphere 5.0: 8Gb/s Fibre Channel SANs with HP ProLiant DL380 G7 Servers and HP 3PAR Utility Storage

Announcing the new Cisco Unified Computing System (UCS) B-Series Blade Servers with Emulex OneConnect 10GbE Adapters (Or as I like to call it, “The Ultimate Data Center Machine”)

Posted March 8th, 2012 by Sonny Singh

“What is it about guys and their obsession with cars?” – If you’re like me, this sarcastic and often repeated rhetorical question is spoken by loved ones (mostly wives) ad nauseum. In fact, I can bet there has been some semblance of this question spoken to us guys since the dawn of time. Example: Cleopatra to Julius Caesar – “Really Gaius? What’s so big about a carbon-fiber war chariot with dual suspension coil-overs that warrant you trading 100 of my prized royal pet crocodiles for one?”

Now, don’t get me wrong, I know a “few” wives that not only appreciate the aesthetics of say, a brand new BMW M3, but also admire its razor-sharp handling and super-tuned 4.0-liter V8 powerplant, delivering a whopping 414 horsepower and a maximum torque of 295 pound-feet at 3,900 rpm. Oh, and let’s not forget about a surprisingly practical interior with fully functional rear seats.

Unfortunately in my case, the issue therein lies with my wife not being able differentiate between a “regular” BMW 3-Series and a “not so regular, super-cool” BMW M3. In her highly rationale mind, the roughly $25K delta between both models is asinine and only a fool would pay that sort of money. After all, both cars look exactly the same right? Two Doors, four wheels, leather seats, sunroof and a BMW emblem (oh, let’s not forget the surprisingly practical interior and fully functional rear seats in “both” models).

You see where I am going with this? I am one of those fools. Guys tend to live by the simple credo when it comes to automobiles – Bigger, better and faster. Unfortunately, this thought process often causes us to make purchasing decisions that don’t embody the notion of practicing judicious financial responsibility. At the end of the day, God bless my wife because I am the first one to admit that if it weren’t for her, I would be driving a $370K Lamborghini Murcielago but also living in it…not exactly conducive to raising a family.

Speaking of bigger, better and faster, the real agenda at hand doesn’t revolve around my obsession with high-performance automobiles. Where the rubber meets the road is really about Intel’s recent announcement of its next-generation Xeon (codename: Romley) processors. Intel has deemed the Xeon E5-2600 platform as “the heart of a next-generation data center.” As the first server processor with integrated PCIe, the E5-2600 platform will increase I/O performance for VoIP, storage/networking and reduce system power and lower I/O latency. Compared to previous-generation Xeon servers, the E5-2600 is a powerhouse platform for 10Gb Ethernet (10GbE) I/O solutions and will rapidly increase 10GbE adoption. In fact, Intel has specifically stated that the proper I/O solution for these new servers is 10GbE.

With the launch of Intel’s new Xeon platform, Cisco and Emulex have also introduced their third-generation solutions which take advantage of the new features delivered by Intel and combining the latest innovations from Cisco and Emulex.

The Emulex OneConnect Fibre Channel over Ethernet (FCoE) Universal Converged Network Adapters (UCNAs) support the new Cisco Unified Computing System (UCS) B-Series Blade Servers. The new Cisco UCS CNA M73KR-E Emulex Converged Network Adapter (CNAM73-KR-E) supports a common 10GbE infrastructure by providing high-performance FCoE mezzanine card that consolidates traffic for networking and FCoE storage for Cisco UCS B-Series Blade Servers giving customers one platform for storage and networking connectivity. Emulex FCoE connectivity enhances the Cisco Unified Fabric and UCS by reducing network complexity, installation and maintenance costs, while optimizing performance and ensuring field-proven reliability for enterprise data centers.

What’s cool about Emulex’s OneConnect technology is that it is based on decades of proven Fibre Chanel heritage and the customized adapter for the new Cisco UCS B-Series blade servers provide customers not only the flexibility of FCoE, with the option of iSCSI offload support, they also can take advantage of Emulex’s demonstrated high-performance, reliability and compatibility with the Cisco Unified Fabric.

Furthermore, Cisco UCS B-Series Blade Servers, leveraging the customized Emulex CNAM73-KR-E mezzanine card, significantly reduce end users’ capital and operating expenses with a converged network fabrics and integrated systems management capabilities. In addition, these servers work with virtualized and non-virtualized applications to increase performance, energy efficiency and flexibility.

The easiest way to look at the synergistic relationship between Intel, Cisco and Emulex is by referencing my little backstory on the BMW M3. When you couple the new Intel Xeon next-gen processors with the Cisco UCS B-Series blade servers and the Emulex 10GbE UCNAs, you truly get a single platform providing powerful, nimble, high-performance capabilites for storage and networking connectivity. I like to call it the “Ultimate Data Center Machine.” Cheesy? I think not.

For all my grandiose talk, I am humbled to say I currently drive a VW minivan. Does it have 460+ HP and a maximum torque of 295 pound-feet? Not even close. However, will swallowing my pride by driving a minivan for few years earn brownie points with my wife and open the door to eventually getting my coveted BMW M3? ABSOLUTELY.

And don’t forget, please visit our Cisco landing page for everything you need to know about Emulex solutions optimized for Cisco environments.

Emulex I/O Connectivity Solutions Optimized for New Intel Xeon Processor E5-2600-based Server Platforms

Posted March 6th, 2012 by Emulex Labs Team

Author: Judi Uttal

Year after year, the insatiable demand for computing and networking increases as virtualization, mission-critical applications and emerging cloud architectures punish existing networks and CPU processing with the need for greater I/O. The good news is that the new Intel Xeon E5-2600 processor (codenamed Romley) dramatically improves server processing power, while reducing latency. The really good news is that Emulex connectivity and management solutions fully complement this new Intel technology, maximizing the return on investment (ROI) for data centers worldwide.

With today’s announcement from Intel, Emulex I/O connectivity solutions are optimized for the new Intel Xeon E5-2600-based processor, here are some details and a bevy of resources on how Emulex connects with Intel solutions.

Emulex OneConnect® Universal Converged Network Adapters (UCNAs)
Next-generation servers rely on 10Gb Ethernet (10GbE) for balanced performance to support virtualization, cloud computing and mission-critical applications. Data centers can standardize on Emulex for all 10GbE form factors with a common driver model and management now available from tier one server vendors. Qualified for several new servers based on Intel E5-2600 being announced this week, Emulex OneConnect UCNAs include options for LAN on Motherboard (LOM), custom form factor blade and standard PCIe option cards. In fact, Emulex delivers the only LOM that provides enterprise-proven convergence, with protocol offload for TCP/IP, including TCP Chimney Offload, Fibre Channel over Ethernet (FCoE), Internet Small Computer System Interface (iSCSI) and Network Attached Storage (NAS), delivered through vEngine™ technology. With over two million ports shipped, Emulex OneConnect UCNAs deliver market-leading performance for virtualized environments with key technologies, such as SR-IOV and Universal Multi-Channel (UMC), as well as the only VMware vCenter plug-in for adapter management from the VMware console via OneCommand® Manager.

Emulex LightPulse® Fibre Channel Host Bus Adapters (HBAs)
With over 11 million ports shipped across ten generations of Fibre Channel HBA products, Emulex LightPulse HBAs are widely used for storage area network (SAN) connectivity. To fully optimize a next-generation solution, servers using the new Intel Xeon E5-2600 processor need maximum I/O capacity to support high bandwidth applications such as cloud computing, data warehousing and streaming video. Increased bandwidth is also needed for server virtualization, which consolidates I/O from multiple virtual machines (VMs) to the host’s data path. The Emulex LightPulse LPe16000 family of 16GFC HBAs delivers a record breaking over 1 million IOPS on a single channel and is the only HBA that supports T10-PI data integrity offload for protection against costly silent data corruption events that can cause data loss. In addition to increased I/O rates, the LPe16000 provides reduced response times and better CPU efficiency to maximize server resources, with 3x better IOPS performance per watt through its GreenState™ technology, fully complementing the new server technology. The OneCommand Manager and vCenter plug-in management applications work across all protocols on both Emulex Fibre Channel HBAs and OneConnect UCNAs for single-pane-of-glass management. Boasting a clean sweep of design wins across all tier one OEMs, the Emulex 16GFC HBA is poised to be the premier provider of SAN connectivity for next-generation servers. Look to Emulex to be the first to support PCIe Gen3 with new design wins.

Emulex Pilot Integrated Baseboard Management Controller (iBMC)
Intel has selected Emulex’s Pilot 3 iBMCs as the single management controller embedded in the new Intel Server Boards and Systems supporting the Intel Xeon E5 processor family. With an advanced triple-core architecture, Pilot 3 iBMC delivers enhancements geared at the new Intel processor, offering faster speeds, support for more ports, additional sensors and reduced power requirements. Emulex Pilot 3 iBMC offers the industry leading Clear KVM™ technology for superior remote console operation and high-speed remote storage support. With OEM design wins across premier companies, with new announcements coming, Emulex Pilot iBMC is considered an optimum solution across all baseboard management controller products.

Intel and Emulex
Virtualization, cloud computing and mission-critical applications are the driving forces behind many data center initiatives. With the need to increase capacity while reducing operational expenses (OpEx), IT management looks for ways to use limited capital expenditures to get the maximum benefit. Fortunately, next-generation technologies from Intel and Emulex deliver a quantum shift in server performance and energy efficiency. However, selecting the best server with the new Intel Xeon E5-2600 processor is not an easy task with so many variables to analyze. Fortunately, selecting Emulex as the connectivity provider for your new server is an easy task. It is really the best solution available!

Visit emulex.com/intelxeon.

Windows 8: One Bite at a Time

Posted February 29th, 2012 by Emulex Labs Team

Author: Jean-Yves Chevallier

Every four years or so, Microsoft serves up a major new version of their Windows Server operating system. After Windows Server 2003 and 2008, we are now headed to … Windows Server 2012? Or Windows Server 8? Bets are on the latter. The news is, Microsoft is rolling out the Beta version on Feb. 29.

The Windows 8 cycle balances support for existing computer models (desktop client, corporate and departmental file and application servers) with an aggressive redesign for the emerging “cloud-based” model … and evil intentions of leapfrogging and marginalizing arch-rival VMware. Is this just window-dressing or is there substance there?

Client-side first: the new Windows 8 “metro” look simplifies user interfaces for smaller touchscreens. Think tablet. Think smart phones … Think of devices that haven’t made it out of the labs yet. Thankfully, you can still use your existing GUI (more cluttered, but so much easier to use on your laptop).

The most exciting aspect is on the server side, where Microsoft is bravely tackling the key challenges of the emerging cloud model, where your applications and data will be served 24/7 from unknown locations across the globe. They could possibly be co-located with your competitor, to whichever user device you happen to be working with at the time:

  • Keeping your stuff separate from everyone else’s: virtualizing your desktop, your application environment, your network. Virtual interfaces (SR-IOV virtual functions, synthetic HBAs), multi-tenancy with IP address and virtual LAN (VLAN) mobility, offer a sophisticated infrastructure … and a steep learning curve. Emulex will continue to publish a constant stream of online seminars, white papers, and blogs to help you select and implement the right solutions for you.
  • Scaling up and out to absorb workloads and traffic: more processors, offloaded data transfers, offloaded network operations (TCP and IP checksums, LSO, LRO ), traffic prioritization (with enticing names such as PFC, ETS, DCB, DCTCP and QCN ),
  • Acceleration techniques from SR-IOV (again) to RDMA and SMB-Direct, and of course bigger pipes with 10Gb Ethernet and 16Gb Fibre Channel, (both technologies Emulex happens to dominate) and more.

So how do you implement Windows 8 and use it to realize the benefits of the cloud? You do it in the same way you eat an elephant (one bite at a time). Windows 8 is expected to reach general availability in October 2012. We have time to prepare and cook the elephant, and you don’t have to eat everything in one serving. Keep an eye out for information from Emulex – your Windows 8 chef!


¹LSO = Large Segment Offload (breaks down large outgoing packets for easier handling by switches and routers)
LRO = Large Receive Offload (aggregates multiple received packets to reduce interrupt overhead)
²PFC = Priority Flow Control. ETS = Enhanced Transmission Selection. DCB = Data Center Bridging, DCTCP = Data Center Transmission Control Protocol, QCN = Quantized Congestion Notification (Together, these technologies prioritize traffic and regulate its flow through the network).

ESG Labs conducts hands-on testing of Emulex 16G Fibre Channel HBAs, and why you should care

Posted February 13th, 2012 by Barbara Porter

Enterprise Strategy Group (ESG) just posted a new report which documents ESG Lab’s hands-on testing of Emulex LightPulse ® 16G Fibre Channel (16GFC) Host Bus Adapters (HBAs), and explores the HBA’s ability to improve virtualization efficiency and increase performance in an 8Gb or 16Gb environment. The report also covers the ease of management and simplicity of deployment of the LPe16000 series.

According to ESG research, Fibre Channel is still the primary storage technology used to support virtualized server environments. When asked to name the factors preventing organizations from using server virtualization more pervasively, two of the top three responses were lack of budget and performance concerns. A Fibre Channel HBA that could enhance performance while reducing latency without having to rip and replace existing SANs would be a compelling proposition.

Here are a few of the highlights of the report:

  • ESG Lab was able to deploy and manage a LPe16002 HBA into an existing SAN environment side by side with multiple generations of Emulex HBAs using Emulex OneCommand™ Manager for a single consistent point of management.
  • ESG Lab confirmed that Emulex has developed a robust and very full-featured vCenter plug-in that provides all the functionality of OneCommand Manager, including the distribution of mass firmware updates directly from the vCenter console.
  • The LPe16000 series of HBAs run the same drivers as previous-generation Emulex adapters, simplifying management and maintenance.
  • The performance of the LPe16002 HBA tested by ESG Lab was particularly impressive, driving more than five times the OLTP IOPS of its 8GFC predecessor. Even more impressive was the fact that this increase in performance came with a 50% decrease in latency.

ESG Lab’s report on the new Emulex 16GFC adapters echoes the high-performance, reliability and management functionality that makes our adapters the clear choice for the toughest virtualized, cloud and mission critical deployments. To view the entire report click here.

Big Data Solutions at Emulex

Posted February 10th, 2012 by John Cedillo

I’m sure we’ve all heard some of these staggering internet and data statics that are making the rounds, right? Something along the lines of…

  • 48 hours of video are uploaded to YouTube every minute (source)
  • 8 trillion text messages were sent in 2011 (source)
  • An estimated 100 billion photos have been posted on Facebook (source)
  • Twitter logs over 250 million tweets per day (source)
  • We create 2.5 quintillion (2.5 x 10^18 ) bytes of data on a daily basis (source)
  • 90% of the data in existence today has been created in the last two years (source)

For a little perspective, IDC has said that it has taken almost 60 years for disk drives to reach 1.7 zettabytes (ZB) of data in the Storage Universe and they expect that to almost quintuple to more than 8ZB by 2015. They also predict that 90% of this new data will be video and pictures¹.

The sheer scale of this boom in data growth, and the sources driving it, such as consumer participation in the web, social media, mobile applications, credit card and banking transactions, and high frequency trading, just to name a few, show no signs of slowing down.

In the traditional sense, “Big Data” has been used to describe massive amounts of data controlled and analyzed by huge organizations like Google. Below that echelon of organizations, “Big Data” is a relative term, proportionate to the size of an organization. Regardless of size, sector, or vertical, the exponential growth of data and focus on data analytics has prompted companies to adopt Hadoop – to uncover new and valuable information from unstructured data sets, and turning that into a competitive advantage for their business.

The proliferation of data, and data sources, is exactly why this topic is growing in popularity – it relates to the core of the emerging models that tackle the analysis of this massive data growth. Emulex OneConnect® 10Gb Ethernet (10GbE) Network Adapters eliminates the performance bottleneck imposed on Hadoop clusters by traditional 1GbE networks, unlocking the full potential of today’s servers. When combined with OneCommand™ Vision, Emulex is able to provide administrators with the most efficient and scalable solution for maximizing the performance of their Hadoop clusters.

Our new Big Data Solution Site arms you with the resources you need as you move forward with I/O connectivity and management deployments to manage massive amounts of data in your organization. We offer a whitepaper detailing OneConnect 10GbE adapter performance improvements over 1GbE connections as well as a whitepaper on OneCommand Vision and how I/O management is critical to Big Data solutions. Links to the latest industry blogs also provide up-to-date information on what is happening in the industry.

You can also visit us in Booth #522 at the upcoming Cloud Connect conference in Santa Clara next week, and learn first-hand what Emulex offers in the way of cloud computing and Big Data solutions.


¹Source: IDC: IDC Predictions 2012: Competing for 2020 (#231720, December 2011)

«Older Posts