Emulex Blog: Emulex Labs

Cisco Propels Cloud-Ready Switching Portfolio Into Next Frontier of Networking Speed and Simplicity

Posted February 9th, 2012 by Sonny Singh

A few years back, the world as I knew it got rocked to the core. No, I didn’t lose a loved one to a sudden and unexpected illness, nor did I find out my wife was leaving me for the pool guy (even though I don’t have a pool, let alone a jacuzzi) but you get the point. In my mind, this catastrophe was on par with something just as cataclysmic … I lost my entire library of digital music due to a dead hard-drive. Over 10,000 of my beloved songs ranging from “Hungry like the Wolf” by Duran Duran to “Nothing But a G Thang” by Dr. Dre vanished within the blink of any eye. Having worked in the data storage industry for years, I had no one to blame but myself for not diligently following a data backup policy. It took me about two months of painstaking labor to rebuild my library from scratch … oh well, live and learn right? Well, I can tell you that was the first and last time it ever happened because I discovered a little thing called “The Cloud.”

So what is the cloud? In the most simplistic of terms, the cloud is as old and simple as the Internet itself. The cloud is really just about accessing storage or software remotely from a computer via the Internet. However, as simple as it may seem, the cloud is big business.

Keep in mind, when you start talking about the cloud being deployed in data center environments, the benefits it provides increase tenfold. This leads me to talking about Cisco’s recent announcement on February 1, which further validated their commitment to help customers build intelligent infrastructures ready to support cloud-based environments.

According to the Cisco® Cloud Index, more than 50 percent of computing workloads in data centers will be cloud-based by 2014, and global cloud traffic will grow more than 12 times by 2015, to 1.6 zettabytes per year – the equivalent of more than four days of business-class video for every person on Earth. Based on these trends, Cisco announced it updated its switching portfolio with 100 Gigabit Ethernet (100GE) and 40GE capabilities, the next speed limits for networking. These new capabilities provide a holistic architectural approach across campus, data center and service provider environments, and give customers a significant advantage while addressing emerging trends like cloud, video, mobility, and the proliferation of 10Gb Ethernet (10GbE).

With this announcement, Cisco, which already offers one of the industry’s broadest switching portfolio, becomes a key provider of 1/10/40/100GbE and converged networking switching solutions. Supporting the transition to higher networking speeds, in addition to Cisco’s recently announced CloudVerse solution, is a key element of Cisco’s strategy to help its customers meet the demands of emerging cloud computing trends and the resulting data deluge.

Cisco is delivering investment protection with 40GbE performance options to its Catalyst 6500 switching line, and 40 GE and 100 GE capabilities in its Nexus 7000 portfolio for interconnecting data centers to service providers. To expand its campus aggregation and data center top of rack switching, Cisco also announced two new fixed-configuration platforms that provide high-density 10GbE switching. Making it easy to use this capacity, Cisco announced simpler network virtualization functionality for its Catalyst 6500, 4500 and Aggregation Services Router (ASR) 1000 product lines with a new technology called Easy Virtual Network. They also added scalable virtual services with a new Nexus 1010-X appliance for the data center. Together, these enhancements will help businesses scale their networks, simplify operations, and protect their existing Cisco switching investments – many of which have been deployed for a decade or longer. Cisco recently reached switching milestones of 10 million deployed 10GbE switched ports since 2002, and 23,000 NX-OS customers, the operating system that supports Cisco’s entire data center portfolio.

And where does Emulex fit into all of this? Emulex has worked closely with Cisco to ensure that our Emulex OneConnect 10GbE LAN on motherboard (LOM) and 10GbE Universal Converged Network Adapters (UCNAs) are fully compatible and maximize IT managers’ investments in their Nexus and MDS infrastructure. So, whether you or your customer are looking for an answer on a Massively Scalable Data Center, Private/Hybrid Cloud, Big Data, High Frequency Trading or a traditional Data Center, Cisco and Emulex have the solution for you.

In closing, my updated library of digital music has grown to about 15,000 songs since that traumatic incident a few years back. Granted, now that I have three boys under the age of five, the spectrum of musical tastes ranges from everything Disney to Justin Bieber (God help me). However, at the end of the day, I can go to bed in peace with the sublime knowledge that if one of my twin boys decides to projectile vomit on my laptop (and also manages to hit my external hard drive three feet away), I have nothing to worry about. Sure, my $2,500 Mac will be fried into oblivion and I’ll need to bust out my carpet cleaner for the umpteenth time.

What matters most is I’ll be dreaming peacefully about my library of digital songs being nice and safe … nestled in the cloud, with an army of server storage standing sentry over them every night ready to do battle with a bad hard drive (or vomit for that matter).

For more information on Cisco’s recent announcement please visit:

http://newsroom.cisco.com/press-release-content?type=webcontent&articleId=657081
http://blogs.cisco.com/news/cisco-switching-leadership-and-innovations/

Do you need to outsource your storage but have reservations? Manage it with OneCommand™ Vision!

Posted February 8th, 2012 by Emulex Labs Team

Author: Kati Bell

Here’s an interesting dilemma out there in the real world: more and more data centers are outsourcing their entire on-premise storage infrastructure to a third-party. In talking with some industry folks, it turns out that this is a scary proposition. What’s needed is a fresh approach to storage performance monitoring, one that lives in the client’s domain and not the third-party’s domain. Emulex OneCommand Vision does just this, allowing the benefits of outsourced storage, while maintaining control. Check out what our customers are doing.

To enable our customers to have increased control over external or internal storage operations, we have recently added cutting edge LUN service level agreement (SLA) management to OneCommand Vision version 2.1, allowing administrators to establish LUN SLAs that consider throughput, I/O operations per second (IOPS), path availability and I/O latency. These new LUN SLA enforcement capabilities enable monitoring and reporting services tailored for private or public cloud deployments.

OneCommand Vision 2.1 also delivers support for IBM AIX Power and role-based management and GUI data export, allowing users to save data in PDF and CSV formats. New intelligent event decoding has also been added, starting with support for SCSI check condition and RSCN events.

In addition to support for IBM AIX Power, Vision 2.1 also includes expanded OS and hypervisor support for the following
Hypervisor Enabled OS:

  • IBM AIX
  • Linux KVM
  • Oracle Solaris
  • VMware ESX
  • Window Hyper-V

Non-Hypervisor Enabled OS:

  • Linux (Redhat, SuSE, Oracle)
  • Windows

To learn more about OneCommand Vision, please visit us at Emulex.com/vision and let us know what you think of these new enhancements. Or, join our webcast on 2/15 on “OneCommand Vision 2.1 : Cutting edge LUN SLAs, AIX on PowerPC and flexible data export options put YOU in control”, to learn more!

Emulex OneConnect Twisted Sisters: Announcing the new OCe11102-T Adapters

Posted February 7th, 2012 by Emulex Labs Team

Emulex has announced availability of the “twisted sisters” of the OneConnect 10 Gigabit Ethernet (10GbE) product family. The two adapters are the OCe11102-NT network adapter and the OCe11102-IT iSCSI adapter with full protocol offload.

These adapters support 10GbE connectivity using industry-standard Cat 6 and Cat 6a twisted pair cabling, and standards-based 10GBASE-T technology. There are several key benefits that make these adapters a great solution for many data centers:

  • Work with commonly used twisted pair Cat 6 and Cat 6a cabling
  • Lower costs when compared to optic cabling
  • Longer cable runs when compared to direct attach copper (DAC) cabling
  • Backward compatible with 1GbE networks to “future-proof” servers during the migration to 10GbE
  • Based on market-leading, award-winning Emulex OneConnect 10GbE network technology

Detailed product information has been posted on the Emulex web site on our landing pages: OCe11102-NT and OCe11102-IT. Contact your Emulex sales representative, favorite reseller or distributor to learn more.

16G Fibre Channel: Bigger and Badder FC For Virtualization, Cloud and Database Applications

Posted January 20th, 2012 by Barbara Porter

Most virtualization deployments rely on storage area networks (SANs) for flexible shared storage solutions to meet mobility, performance, scalability and efficiency requirements. As many data centers take the next steps in virtualizing big I/O applications, like databases, and move to more scalable private clouds, storage networking has become the primary bottleneck for Quality of Service (QoS) and scalability.

The new Emulex LightPulse16G Fibre Channel (16GFC) Host Bus Adapters (HBAs) fix that bottleneck, enabling the best QoS for the highest virtual machine (VM) density with the fewest ports and cables and the lowest power footprint.

Additionally, the entire SAN fabric benefits from higher availability and reduced power requirements leveraging a faster HBA. Because of better performance as well as streamlined management and backward compatibility, Emulex 16GFC HBAs is the best solution for virtualized environments. Here is what you can expect when upgrading to Emulex 16GFC HBAs:

  • 5x the IOPS
  • Twice the data throughput
  • Cuts application I/O response time in half
  • Up to 4x the IOPS for typical 4K/8K I/O block database applications
  • 3x the IOPS performance per watt
  • Maximum VM density with increased N_Port ID Virtualization (NPIV) virtual ports (vPorts)
  • True cloud scalability, with support for up to 255 virtual functions and 1024 MSI-X and 8192 logins and open exchanges for maximum VM density—up to 4x more than other 16GFC adapters
  • Unmatched native manageability with Emulex OneCommand Manager for VMware vCenter – enables adapter management directly from the vCenter console, delivering 2x adapter management functionality and taking half the time to install and manage compared to other adapters
  • End-to-end data integrity with BlockGuard™ hardware offload – supports
    the T10 Protection Information (T10-PI) standard to protect against silent
    data corruption, without the 30-40% performance tax incurred by other
    firmware-based T10-PI solutions

If you’d like to learn more about 16GFC technology, join our webcast next week that will be presented by industry veteran and analyst, Seamus Crehan, principal of Crehan Research. The webcast will cover drivers for the 16GFC transition and the impact of virtualization and upcoming server refresh cycles on the overall SAN market. Shaun Walsh, vice president of marketing at Emulex, will also participate in the webcast. Click here to register for the webcast on January 25th at 9am. Come armed with your questions!

Click here to Download the paper

*Performance numbers compared to Emulex 8GFC adapters.

The 2012 Technology Cycle

Posted January 18th, 2012 by Emulex Labs Team

Author: Jean-Yves Chevallier

New, disruptive technology that is invented daily cannot be absorbed instantly by organizations. To facilitate digestion, the computing industry is increasingly bundling it into coherent technology cycles. One such new cycle is starting now and will unfold through 2012.

What ‘s the new stuff?
The 2012 server technology bundle includes four big components: the Intel Romley generation of servers (eight cores, 20 percent performance increase), 10Gb Ethernet (10GbE) on motherboards, 16Gb Fibre Channel (16GFC) adapters and networks, and new operating systems such as Windows 8. Note that the last two show how resilient good enterprise technology can be: wasn’t the death of Fibre Channel heralded several years ago? Wasn’t the Microsoft OS marked for extinction?

The good news is, these technologies are not just dropped randomly by individual vendors. This cycle has been two years in the making, with vendor cooperation ranging from conceptual designs to plugfests and comprehensive integration cycles. As an example, Emulex communication processors and storage adapters have been tortured by server vendors on prototype motherboards and early chassis, and thoroughly abused by operating system vendors (while our own labs performed similar indignities on early servers, blade and rack-top switches, and operating system builds).

The results have started trickling out, and will continue to do so in the coming months. Emulex fired the first salvo with the first 16GFC adapters on this galaxy, announced in October 2011. Intel Romley servers by major vendors will be unveiled in the coming quarter, bringing with them built-in 10Gb connectivity. Bookending the cycle, Microsoft Windows 8 was first unveiled at the September 2011 BUILD conference, and will be in general release at a still-confidential date in 2012.

So what? Why should you care?
Two words: virtualization, cloud. Virtualization, once a way to jettison old print servers, has evolved into a powerful architecture, and the 2012 cycle brings it to the next level. Words like Synthetic HBAs, Single Root I/O Virtualization (SR-IOV), and Storage Spaces, will become everyday words for organizing your production environment. Possibly more important , the cloud, once the province of gurus and jokers, will be supported by a wide set of rapidly converging technologies, from encapsulation to flow control and, yes, virtualization again.

Go full-tilt virtual, reach for the cloud: 2012 is here!

Announcing the Emulex Extra Program—Thanks for your feedback!

Posted December 7th, 2011 by Emulex Labs Team

We are thrilled to announce our new Emulex Extra Program which is designed to provide end users with access to new tools, training and support, including free technical training and certification. Emulex Extra is the direct result of a survey of nearly 1,000 customers which found that 88% of our end users ranked technical support as a top priority, and more than half felt that training was also a priority. This program is a direct result of your feedback! Join us on December 14, at 9:00 a.m. PST/12 p.m EST for a webinar to learn more. Click here to register now.

Here are a few program highlights:

  1. Emulex Expert Certification: Become an Emulex Expert through our free Emulex Extra certification training. Emulex top-rated computer-based training (CBT) can make you an expert on deploying, managing and configuring Emulex OneConnect™ 10Gb Ethernet (10GbE) Universal Converged Network Adapters (UCNAs) and LightPulse® Fibre Channel Host Bus Adapters (HBAs), as well as our OneCommand™ management software.
  2. Extended Support: As an Emulex Extra member, you are entitled to Emulex Technical Support (24×7). The Emulex Support team has hundreds of years of experience and will personally help you solve your networking issues. As an Emulex Expert Member, you will be able to log and monitor your own support cases, using the Emulex support database.
  3. Customer Portal: Access the Emulex interoperability tool, whitepapers and downloads. Emulex interoperability testing combined with our ecosystem partner collaboration, ensures that our products work well in your data center – regardless of the software, server, storage, and switches that you deploy (see our recent blog on how to use the interoperability tool, here). The Emulex interoperability database provides easy access to industry-wide test results, giving you confidence that what you install will “just work.” In addition, Emulex publishes hundreds of whitepapers, videos and webcasts every year. Use our simple online search tool to find your desired resource based on keyword, document, product or partner.

And, for users that already have a “MyEmulex” account, you are already enrolled in the Extra Program.

These are just a few of the program perks. Please join us on December 14 to learn more about this exciting new program. AND visit the Emulex Extra Program landing page to register for the program and to learn how you could win $100.

Questions? Please download the Emulex Extra FAQ. Comments? Email us at: Extra@Emulex.com. And join today!

Technology Interoperability – An Ecosystem Lexicon

Posted December 5th, 2011 by Emulex Labs Team

By John Cedillo, Emulex

IT administrators are faced with an enormous number of hardware and software solutions. In fact, using the simplified diagram below, which is far from being a comprehensive set of choices that exist in designing data centers, there are 42,120 different configuration possibilities – and this doesn’t even begin to consider the various appliances that would connect into the data/storage networks, or the applications that will sit on top of the operating systems. From one end to the other, there is a common reliance on the storage and networking controllers themselves, which is why we at Emulex strive to be the leader in ecosystem interoperability.

Navigating these design considerations is a pervasive challenge – from administrators selecting upgrade paths for their existing data centers, to organizations planning brand new data center installations. Thoughtful and informed planning is an essential step to combat and alleviate the complexity associated with disparate vendor systems within a data center, and eliminate downtime.

With the leadership position that Emulex has taken with pushing new technologies into the mainstream, combined with the unique position that our LightPulse® and OneConnect™ adapters occupy in the data center, it was the ideal time to redesign our interoperability database in an effort to provide our customers with the data they need to make informed storage and networking decisions.

We have created two “entry points” for our interoperability tool.

The first method allows a user to search for product interoperability based on a combination of Speed and Protocol – once selecting a Protocol from a Speed grouping, the user can quickly locate which server and storage OEMs have tested and integrated Emulex adapters into their respective product offerings. Selecting the desired OEM from the Partner column will allow the user to quickly identify the OEM’s hardware family and orderable part number.

The second method allows a user to search for product interoperability based on a specific Partner within grouped product lists – providing a complete listing of all Emulex products that have tested and certified interoperability with that partner’s hardware portfolio. (i.e. HP Servers, Extreme Networks 10GbE Switches, Cisco FC Switches, etc.)

No matter the entry method you choose, we’ve added some informational high points that we think will be useful and informative.

We’re proud of the enhancements that we put into our interoperability tool, and we feel that these changes will be helpful to you. That doesn’t mean, however, that we can’t continue to improve upon these changes, and it is in that spirit, that we ask for your comments or suggestions for improvement. Please email us at InteropMatrix@Emulex.com – we look forward to hearing your feedback.

If you are interested in learning how to go about working with Emulex on interoperability and certification testing, please contact me at John.Cedillo@Emulex.com – we are eager to work with ecosystem partners!

What is OneConnect UMC and why should I care?

Posted November 21st, 2011 by Emulex Labs Team

It’s taken a while, but we’re seeing the transition to 10 Gigabit Ethernet (10GbE) reach critical mass. Next-generation, high-performance servers are enabling higher virtualization ratios and driving demand for increased bandwidth. The economy has been tough on budgets, but lower prices for adapters and switches are making 10GbE much more affordable.

With basic economics of price and demand working in the right direction, there are critical questions on how to make the best use of 10x bandwidth:

  • How can I conform with best practices for server virtualization?
  • How can I ensure critical virtual machines get needed bandwidth?
  • How can I optimize network and storage traffic that are sharing the same infrastructure?

Emulex has introduced Universal Multi-Channel (UMC) as a key technology that works with OneConnect OCe11102 10GbE adapters to answer all of these questions. With UMC, each physical port can be partitioned into four separate Network Interface Cards (NICs) functions with a unique MAC address and virtual LAN (VLAN) assignment. There’s also the option to support iSCSI or Fibre Channel over Ethernet (FCoE) traffic with full protocol offload on one of the functions.

Instead of installing eight 1GbE NICs with all the related cables and switch ports, you can use one dual-port OCe11102 adapter with eight NIC functions and a default bandwidth of 2.5Gb per function. That gives you the flexibility and scalability to truly leverage your investment in 10GE.

Want to know more?

We’ve posted a new installment of our Solution Implementer’s Series with details on how to deploy UMC for best results: Deployment Guide: UMC for OneConnect OCe11102 Adapters.

You can also view a video presentation below by Saurabh Sureka, a technology expert from our engineering team, that highlights the capabilities and benefits that UMC provides.

Emulex®, Intel® and IBM® report on UEFI

Posted November 16th, 2011 by Emulex Labs Team

The Unified Extensible Firmware Interface (UEFI) is a specification that defines a software interface between an operating system and platform firmware. UEFI is an important standard used everyday around the globe. Emulex has collaborated with Intel for nearly ten years on driver development for their Host Bus Adapters (HBAs) in a Fibre Channel Storage Area Network (SAN) using this standard as it has evolved. Intel’s recent Intel Technology Journal features UEFI articles by Emulex, IBM and Intel, that discuss background, architecture and development tips for this UEFI.

Emulex offers UEFI development insights

Emulex engineers Terry Kirch and Al Rickey, provide their UEFI development insights for this Intel publication. Beginning with the Itanium® servers, Emulex developed the “legacy” (X86) driver. With Emulex’s expertise adding to the driver development process, the X86 boot process quickly evolved to EFI, with monolithic driver architecture developed that combined Block I/O, SCSI I/O and SCSI pass-through direct. As EFI evolved, Emulex took a lead role in replacing the X86 driver with the EFI (1.10) driver. Most recently, Emulex developed the HII driver, based on UEFI 2.x, for its HBAs.

Read this interesting article by Al and Terry, beginning on page 57, for some historical background information, mixed with current development tips. It will change the way you look at your driver installation the next time you install an Emulex adapter.

Supercomputing 2011 Sneak Peek: Showcasing 10Gb Ethernet in Hadoop Environments

Posted November 11th, 2011 by Emulex Labs Team

Next week, at Supercomputing, we’ve got a first-hand look at a joint solution we’ve been working with our partner AMAX on the benefits of 10Gb Ethernet (10GbE) to Hadoop performance. Speaking of Hadoop, you must be wondering what all the buzz is about this Hadoop ‘thingy’? If you are like me, it seems to pop up at every technology summit or is mentioned in some industry experts’ blog or another. Well, it’s safe to say, Hadoop is indeed not associated with a new type of dance move coming out of some night club in Hollywood (in fact, it was named by the founder after his son’s beloved toy elephant).

Hadoop is a software framework that supports data-intensive distributed applications under a free license and it enables applications to work with thousands of nodes and petabytes of data. What’s really cool about Hadoop is it is designed to scale up from single servers to thousands of machines, each offering local computation and storage. Rather than rely on hardware to deliver high-availability, the software itself is designed to detect and handle failures at the application layer, therefore delivering a highly-available service on top of a cluster of computers, each of which may be prone to failures.

Emulex and AMAX will be demonstrating a Hadoop data cluster configuration that we’ve collaborated on which leverages the AMAX ServMax SX-2208 storage server and Emulex OneConnect OCe11102-N Dual-Port 10GbE Network Interface Cards (NICs). This joint solution will showcase how 10GbE improves Hadoop performance and scalability, opening up new possibilities for Hadoop in traditional and hybrid data centers.

The demonstration will be held in AMAX booth #915 and highlighted in Emulex booth #5205. If you come down and check out the demo, Emulex and AMAX will give you a 2G flashdrive wristband! Wear your wristband at the show and you could win a Starbucks gift card instantly. Everyone who visits the Emulex booth will be entered to win an AppleTV and will walk away with a Ninja Nic Stress Reliever.

As you can see, there is something for everyone at Supercomputing this year. We will be chronicling the week with pictures, videos and live tweets and will share them on our Facebook page as much as possible. Hope to seeing you all in Seattle!

Newer Posts»«Older Posts