Not logged in » Login
Sep 27 2018

Fujitsu Server Components: The Right Solution for Every Scenario


Every year, vendors and customers alike flock to events like Las Vegas's CES or Berlin's IFA, to show off and learn about the latest trends in consumer electronics. For those among us who are on a quest to build the perfect home entertainment system, both shows are a must-see – only there will you find the adequate audio and video components, storage media, networking cables, and accessories that will elevate you to a whole new level the next time you listen to that overwhelming new 7.1 remaster of your favorite album or watch this year's blockbusters in full and glorious 4K resolution.

Sooner or later, however, even the most ardent AV perfectionists may have to give in to external restrictions. After all, what good are a 65-inch flat screen and Dolby Atmos sound system if your living room isn't big enough to enjoy the new system to the fullest, or if you live in an apartment block where neighbors with small children might feel forced to listen in on the latest release from Beyoncé and Jay-Z or your copy of "Jurassic World: Fallen Kingdom"? So eventually they end up purchasing smaller, not entirely flawless equipment and patiently wait for the next advancements in AV technology.

CIOs, CISOs and IT administrators usually share the same appetite for perfection. But unlike buyers of home entertainment systems, they can't wait forever – they have to set up the optimal infrastructure now, without any delay or compromise, and if they don't get around to it asap, their jobs may be at risk. That's why every "IT person" worth their salary will keep a close watch on both the quality of the servers and server components they deploy and their adaptability to different usage scenarios, in order to not only provide maximum performance and meet strict SLAs, but also to avoid unnecessary outages and premature wear and tear.

Components – the Key to Performance and Flexibility
As we have pointed out earlier this year, the idea that "quality matters" has shaped Fujitsu's production process since we first started building x86-based servers in 1994. One way to ensure that our systems can be held to (and meet) this high standard is through the use of first-class components that reliably deliver top-notch performance. But as noted above in our consumer electronics example, sheer performance is rarely if ever the single pivotal criterion in a purchase decision. Put bluntly, an extremely powerful system may turn out to be hopelessly oversized, expensive and costly to operate. A weaker system on the other hand may not provide all desired functionalities. Confronted with two bad options, users will normally start looking for a third way and either opt for another vendor or stick with existing equipment for a couple more years. However, from a sheer customer perspective it would be far better to have a vendor that can provide you with a range of options to choose from, in order to achieve the degree of flexibility your business needs – in other words, one that offers a broad portfolio of tested and qualified building blocks which in turn support a wide variety of application/usage scenarios and fit into diverse infrastructures from small to medium-sized businesses all the way through to large corporations.

For us at Fujitsu, this meant we had to make our systems more flexible and customizable. That's why today, all our PRIMERGY and PRIMEQUEST systems are not only always equipped with the latest state-of-the-art CPUs, GPUs, memory modules, storage or networking gear, but also highly configurable and expandable, in order to accommodate specific individual customer needs. To achieve this, we offer a comprehensive portfolio of server and rack components including:

  • Coprocessors and graphics cards
  • Controllers and PCIe SSDs (e.g. our FUJITSU PRAID EP440i TFM and EP440i TFM SafeStore RAID controllers)
  • Memory modules
  • Extension subsystems
  • Backup and storage units (i.e. tape and RDX drives)
  • Uninterruptible power supplies (UPS) and battery units
  • Switches
  • Rack consoles
  • Power distribution units (PDUs)

While many of these components come from long-term technology partners like Intel and NVIDIA, quite a few are also developed in-house, often in collaboration with specialists. A selection of such components is listed below.


Fig. 1: The FUJITSU PRAID EP440i TFM and EP440i TFM SafeStore RAID controllers have set new speed and data security standards for the operation of RAIDs

A Few Examples
Aside from the occasional tower server, which is typically found in small settings such as workgroup environments, today's IT teams predominantly manage rack servers that reside in compact chassis spanning anywhere between 1 and 10 height units. These servers need a 'home' – and like humans, they prefer one that will provide them with enough room to breathe, an adequate number of electrical connections, sufficient cooling options and a door to keep out unwanted visitors. Fujitsu's PRIMECENTER M1 and PRIMECENTER M2 racks offer all the necessary features – and more: They

  • Will house our standard 19-inch servers and storage products as well as equipment from most other vendors
  • Offer optimized rack structures that support different rack frames, thus allowing for efficient installation of server units, switches, power supplies, and cables and best possible airflow
  • Come with lockable front and rear doors and side panels for theft protection and optional anti-tilt protection
  • May be customized with a multitude of accessories, including rack mounting kits, I/O and KVM switches, UPS solutions, and power distribution units

The PRIMECENTER M1 is available with two sets of rack structures, a symmetric design that's either 600 or 700 mm wide, or a 700 mm asymmetric design that supports lateral cabling. Capacities range from 16 to 42 height units (U) with up to 6 vertical slots for switches or cable routing; the maximum load per U is 20 kg (44 lbs.) and the maximum permitted total weight per 42U rack amounts to 1,000 kg (2,204 lbs.). By contrast, the PRIMECENTER M2 sports two symmetric rack structures – a 600 mm design with 24 or 42 height units or an 800 mm design with 42 or 47 height units. The maximum total weight is limited to 20 kg per U and 1,130 kg per full 47U configuration.

The FUJITSU Server PSWITCH 2048 for its part delivers first-grade connectivity and data transfer rates, simplifies networking and includes a number of key functions required to tackle major virtualization, consolidation and cloud computing projects. As the name implies, this switch features a total of 48 x 10 GbE downlink ports that use standard 10GBASE-T or SFP+ connectors plus six 40 GbE uplink ports with QSFP+ interfaces. This is already a more-than-decent hardware setup; however, what really sets the PSWITCH 2048 apart from the competition is its comprehensive set of advanced (software) features:

  • Auto-discovery ensures that it can be swiftly identified and integrated into a network using pre-defined parameters, without human intervention
  • End Host Mode (EHM) helps to create specific port settings for connecting to a production network, so that admins can set up secured ports without having to worry about aspects such as VLAN mapping or load balancing
  • Data Center Bridging, DCVPN, Edge Virtual Bridging, and FIP Snooping all help to establish, streamline and protect connections between different physical and virtual network layers, thereby supporting large-scale virtualization and software-defined networking

As a 1U low-latency, Layer 2/3 Ethernet switch with redundant, hot-pluggable power supplies, the FUJITSU Server PSWITCH 2048 is ideally suited for Top-of-Rack deployments. Its rich feature set paves the way for a smooth transition towards hyper-converged infrastructures (HCI).


Fig. 2: The FUJITSU Server PSWITCH 2048 delivers first-grade connectivity and data transfer rates while supporting the transition to hyper-converged infrastructures (HCIs)

Fujitsu's Integrated Battery Backup Unit (FJBU) is a viable alternative to conventional uninterruptible power supplies, or UPS's. The latter provide an absolutely critical functionality – the ability to keep a server running in the event of a sudden power outage, in order to enable a graceful shutdown and prevent a major data loss – but their usefulness comes at a price: 'Classic' USP's are typically large, heavy, expensive, and complicated to manage, even more so if they turn into full-blown cabinets, as it frequently happens in large data centers or server farms. The FJBU on the other hand aims to simplify the entire job of providing short-term supply of emergency power and in so doing takes our modular approach – the concept of building servers from interchangeable standard components – one step further. The FJBU has the same size as any of our modular power supplies and can thus be fitted into the same slots to replace a redundant power supply unit (PSU). The new, simpler setup eliminates the need for extra floor space and external wiring. What's more, as an integral part of the system the FJBU draws power from the same AC source as the other components, which drastically cuts standby power consumption and heat emissions. Each backup unit is powered by NiMH batteries that provide a large capacity of 23 Wh (watt-hours), resulting in a backup duration time of five minutes at 280 W. In addition, these batteries are also known for their longevity that easily covers the complete 5-year lifespan of a Fujitsu server, whereas conventional UPS solutions with valve-regulated lead-acid batteries only last about half as long.


Fig. 3: Our FJBU helps to avoid data loss, eliminates the need for extra PSU floor space, and drastically cuts standby power consumption as well as heat emissions

Useful as it is, the FJBU wasn't Fujitsu's first foray into the area of power supplies. That's because PSU's, and by extension cooling solutions, are of critical importance if you're building highly energy-efficient servers, as we have been doing for over a decade. Today, most PRIMERGY servers are equipped with modular PSUs that deliver up to 94% of power efficiency, meaning that only 6% of the transformed energy drawn from the grid are wasted in the form of excess heat etc., as opposed to 20% and more in servers equipped with standard PSU's. This boost in power efficiency also translates into longer product lifecycles and massive cost savings of up to $320 per system over a 5-year period. Moreover, EPA-compliant power supplies also played a pivotal role when we introduced a new 'cooling architecture' in the early noughties. Nowadays known as Cool-safe® Design, it used these PSU's, along with special airflow channels and a mostly cable-less design, to cut down the heat generated by PRIMERGY servers, which essentially led to 20% power savings on average. Since then, we have constantly expanded and refined our cooling concepts and technologies: Cool-safe® ATD, which debuted in 2013, enables PRIMERGY systems to run in data centers where ambient temperatures range from 5 to 40 °C (41 and 104 °F), as opposed to earlier generations, which tolerated between 10 and 35 °C (50 and 95 °F). Select systems are nowadays capable of operating within an even higher temperature range of up to 45 °C (113 °F). As a result, servers can be put into hotter rooms without a need to specifically cool these rooms in advance, enabling customers to build 'chiller-less' data centers. Regarding potential savings on energy and cooling costs, our calculations show that increasing the temperature in a server room by 5 °C can save up to one fifth of the power (and thus cost) previously needed to run the cooling system.

"Paying close attention to detail always pays off in the end" – this principle not only applies to manufacturers of extravagant entertainment systems or luxury cars and their special brand of clients. In fact, it is – or at least should be – even more applicable in regular contexts and everyday settings, such as an organization's data center. In the end, we all benefit from the use of highly customizable, flexible and powerful quality IT gear: Organizations get to work with more reliable systems that can be consolidated into more reliable infrastructures, and can even save money in the process. This easily translates into greater customer satisfaction, and satisfied customers will return to generate a steady stream of revenues for Fujitsu and our channel partners.

Aswin Shankar


About the Author:

Aswin Shankar

Global Product Marketing – Server Products


Comments on this article

No comments yet.

Please Login to leave a comment.


Please login

Please log in with your Fujitsu Partner Account.


» Forgot password

Register now

If you do not have a Fujitsu Partner Account, please register for a new account.

» Register now