OEM Group

Good question but since you’re reading this on Dell’s CIO-focused blog and not Creem, you should assume it is the latter. This is a topic that's been widely discussed across our industry, including this article from Jeff Burt of eWeek.

Grid designs for data center power grid infrastructures need to include careful consideration of uptime and efficiency. Current flow needs to be steady and always available. It should also be cost-effective and “green.”  The most fundamental consideration to be made when planning a new data center is whether to provision it with AC or DC power.  The majority of IT data centers today use  AC power throughout. This is not surprising, given the prevalence of AC power grid infrastructure in centers of global commerce.  A century of AC power grid refinements (i.e. transformers, circuit breakers, fuses, standards) have resulted in a generally accepted view of AC power as dependable. Couple that with these limitations imposed by the perception of DC power grid selection (distances of less than one mile and fewer available hardware choices) and you might conclude that DC power type data centers are rare. This is not true.

There remain a large number of DC powered data centers throughout the world, and new ones continue to be built, driven by telecommunication service expansion across emerging regions. DC power holds advantages for telecommunication networks that relate to always-on “dial tone” service level requirements. Network data flow for audio streaming is circuit-based, rather than packet-based. Slight disruption in sequential read/writes results in a poor user experience. Until recently, 1993 to be exact, AC power was susceptible to a bad thing called “current harmonics,” which can result in overheated neutral circuits and transformers. 

For these reasons, telecommunication companies embraced DC power early in their history of operations. DC power continues to be preferred for telecom networks due to the massive install base already in use, and DC power’s success in delivering incredible “dial tone” service levels -- “If it ain’t broke, don’t fix it,” you might say.

Some other verticals for which DC power is attractive include:  remote and mobile computing (oil rigs, ships, jets), military, and video-content providers. The debate goes on, about the virtues of AC vs. DC power and while the efficiency rating between the two standards has come close to parity over the years, Dell remains a strong advocate of choice in the OEM market. This is why we have chosen to partner with NEI to introduce the NEI E2710, a DC-powered server that meshes with Dell’s DMC systems management console and packs the same performance punch that Dell’s Nehalem-based PowerEdge servers do.

The past few years have seen dramatic enhancements in the efficiency of IT equipment, which has resulted in impressive efficiency gains for AC power equipped servers and storage.  Power supplies now run at 80 percent and higher efficiency levels.  Airflow and cooling design considerations have pushed out bulky component cables, in favor of skinny ones.  Heat dissipation from components inside a  chassis is now carefully considered when product design teams develop new platforms.  Airflow is paramount, low “drag coefficients” enable fans to cool heat producing components with ease.  The fans themselves have been completely redesigned, and are more power efficient themselves.  In some large scale processor intense deployment, servers are “running naked” sans chassis, further eliminating drag coefficients while nestled inside their 19 inch-wide, 42U high racks.  As a result, the power efficiency possible in today’s AC powered data center environments is excellent.  This is fantastic news, both for companies with AC powered data centers and for everyone. 

Now, you have a DC powered server that takes advantage of all of these recent enhancements, and delivers Dell dependability and power.  Use it wisely.

In this interview, Dell's Franklin Flint describes how a customer works with Dell's OEM group, even influencing the design of Dell-branded products for use in their hardware appliances.

Using the Google Search Appliance as an example, Franklin describes how the PowerEdge R710 server includes more network interface cards (NICs), better customization of the LCD display panel, internal flash memory capabilities and other features that make it OEM friendly, or an "appliance-inspired" server platform.

Have a listen:

Recently the buzz in the IT industry has been focused on Software-as-a-Service (SaaS) and the larger concept of “Cloud Computing.” In its most basic form, the concept of SaaS involves a company paying for software as it is used. The software runs on computers someone else owns, and the ongoing maintenance and support of the server infrastructure is handled by that company. It involves essentially leasing the application time for a specific period of time, typically month-to-month. In theory, the value proposition of SaaS is that only paying for computing resources when needed should reduce cost in the long run if an application is used in short spurts.

However, industry analyst firm Gartner recently released the results of a large survey on SaaS implementations still has concerns with broad, cross-enterprise adoption of SaaS. In the study, respondents had some concerns and reservations about specific SaaS solutions.  This suggests that there are some applications where SaaS is a bit ahead of its time. As a Technology Evangelist for Dell’s OEM Group, I don’t see SaaS as a panacea and know there are specific cases where hardware-based appliances should be leveraged or at least strongly considered.

A company needing to deploy a service, application, or IT capability has many options available including SaaS and cloud computing, self-built installations, and an ever-growing selection of appliances. While SaaS is definitely an outstanding choice for many of these applications, it is not for every solution as once imagined. Hardware-based appliances still have a growing role to play and in many instances can overcome many of the challenges that customers like those surveyed by Gartner are experiencing with SaaS.

For example, the third most common issue with SaaS implementations, according to the survey, was reliability and performance to technical specs, something an appliance solution will not have any issues with whatsoever since the solution is designed with dedicated hardware and carefully tested to provide the performance and capabilities for which it was designed. Similarly, the main concern of those surveyed -- functionality for business users -- is addressed with an appliance-based solution since a business need only find the most appropriate appliance for their needs and deploy it in the simple, appliance model (typically a fast and simple installation), and they are up and running with the new solution.

Of course, not every IT service or solution is offered as an appliance, but when they are available, they can be surprisingly simple and effective.

So, what do you think? Are there specific situations where an appliance plays more nicely than SaaS or cloud computing applications?

Using flash storage in computer-based appliances is nothing new. I remember the days of installing PCMCIA flash drives into routers (essentially computers) to update the firmware/OS. That was a great solution back then, but today the cost and complexity of designing a unique single purpose appliance like those old routers is often a significant deterrent.

Dell’s OEM Group has been working with hundreds of customers who have moved their appliance hardware development to Dell and are using commodity based servers for their core products. Until recently they were left with very few options when it came to flash-based storage for those appliances, but with the new appliance server platforms Dell has expanded and improved the flash storage options available to a developer.

Dell’s two main OEM-friendly appliance platforms, the PowerEdge R610 and R710, both offer multiple flash based storage options. They have an internal USB port, an optional internal SD-Card slot, and an optional external SD-Card slot for user removable or replaceable storage solutions.

All three of these options are managed via Dell’s lifecycle management controller (via changes to BIOS setting) to be disabled, enabled, or made bootable – that way an application could activate or deactivate a device on the fly. There are myriad ways these devices could be used in an appliance. For example, the operating system and application could reside on a USB key and no spinning hard drives would be required. One could place diagnostics, store backup configuration settings, hardware licensing keys, factory hard drive images, or whatever fits your solution, on flash devices that are internal to the system. Utilizing these flash storage options reduces support incidents and improves customer experiences while providing a more robust, secure, and reliable platform to the end user.

I have to ask you: if you're using white box servers to power your solution, are you getting this kind of appliance-inspired" design? We'll continue to talk about some of the advantages of our PowerEdge servers and some of the customers that are benefiting from our designs and our close relationship with them.

Since 1995, I’ve had the pleasure to participate in the growth of Dell from a relatively small computer start-up to one of the largest technology providers in the world. What a ride it’s been!

In 2001, I joined Dell's Industry Solutions Group (ISG), typically referred to as the OEM Group within Dell. It’s been ten years since our group began customizing hardware and providing unique services to OEM customers who use standard Dell hardware in their offerings. Last week, my colleague Bruce Eric Anderson, introduced me in a post he wrote on Google's new search appliance. Google is one of our many customers and represent some of the topics that I’ll be writing about here on Inside Enterprise IT.

As a Systems Consultant, I’ve assisted hundreds of product managers, engineers and developers create hardware appliances for their solution. I’ve helped them manage through hardware transitions, and, I’ve informed them about the future of Dell hardware solutions so they could be prepared for all the improvements that come with the advancements of technology.

This past March, Dell released the first of its 11^th generation of PowerEdge servers based on Intel’s XEON 5500 series of CPUs, the PowerEdge R610, R710 and T610. These truly innovative servers, along with the recently released PowerEdge R410 and T410, incorporate many improvements driven specifically by our OEM customers. I have seen several generations of PowerEdge servers marketed as appliances by our customers, and every generation has been consistently more reliable, stable, and feature-appropriate than its predecessor.

We continuously listen to the needs of our customers and use that information to improve the products and services that help them achieve their market goals more effectively and efficiently.

So, I'm happy to be able to share my industry observations with you here and hope that you'll let me know your thoughts and feedback. If you want to reach me directly with questions, you can ping me at franklin_flint at Dell (dot) com.