Enterprise organizations have embraced blade servers for hosting increasingly large sets of mission-critical applications that rely on scaling and availability. The marriage of cloud technology and hardware in hybrid networks has proven to be a valuable infrastructure for reliably supporting these applications.
However, even with the introduction of the cloud, network hardware still consumes large amounts of space and energy. Load balancing has enabled organizations to cost-effectively scale while simultaneously improving performance.
- Dell PowerEdge FX2 Chassis
- HPE ProLiant BL460c
- Lenovo ThinkSystem SN550
- Cisco UCS B200 M5
- Huawei FusionServer E9000
- Fujitsu Primergy BX400 S1 Cloud
- NEC Express5800 Blade Enclosure M
- Dell PowerEdge MX840C
- HPE Synergy 660
- SuperMicro SuperBlade
The benefits of blade servers are the result of their stripped-down architecture, reducing them to the power essentials. Their modular design minimizes physical space and energy requirements. A great many blades can be housed inside a blade enclosure. While a typical rack can house up to 42 units (42U), you can pack many hundreds of blades into a similar space.
Little more than thin circuit boards, each blade server can be dedicated to one application or many. They focus on an element such as compute power at the expense of other elements. A common blade architecture is to group as many CPUs as possible into the available space and leave little room for storage. Some organizations may remove storage altogether – storage, after all, tends to be bulky.
Another benefit of the blade architecture is that it reduces maintenance time, as these servers can be rapidly slid in and out. Large data centers, such as those used by Amazon and Google, operate with huge numbers of blades. Redundancy ensures that the failure of one or more blades has no impact on service. Administrators can go around during a set schedule to replace those that failed.
There are a number of factors to consider when choosing a blade server that all depend on an organization’s business objectives and existing architecture.
These days, vendor consolidation has reduced the options out there. Other vendors have even dropped out of the market, leaving space for HPE and Dell to dominate the market. However, for the purpose of this guide, we have only included two blades each from those vendors. Another six options are covered in the guide.
The FX2 chassis combines the density and efficiencies of blades with the simplicity and cost of rack-based systems. Thanks to its ability to rapidly scale, it’s a popular option for big server farms that want stripped-down compute blades. The Dell PowerEdge FX 2 comes in a 2U rackmount form factor that can be configured to hold 4 half-width, 8 quarter-width sleds or 2 full-width sleds.
This is also a great option for reducing downtime. Because so many blades can be packed together, they can failover to another without impact. It is a good candidate for anyone wanting an entry-level blade infrastructure. But it lacks the storage and compute power of many of the other products in this guide.
The HPE ProLiant BL460c Gen10 was designed for a wide range of configuration and deployment options. The standalone blades are not the most expensive option, but the price tag quickly rises due to the price of its BLc7000 enclosure. Its blade came out just behind Cisco in user review ratings but beats Cisco in terms of the storage options and amount of cores it can support. The ProLiant BL460c can scale up to 26 cores, internal 12 Gb SAS and 2.0TB of HPE DDR4 SmartMemory.
HPE has also made security a focus with the ProLiant BL460c. Major firmware is anchored directly into the architecture, running millions of lines of firmware code before server operating systems boot. It also offers secure recovery, so firmware can be recovered to a good state after compromised code is detected.
It’s a good choice for SMBs and as a general, entry-level business server.
Read our in-depth coverage of the HPE ProLiant BL460c
See user reviews of the HPE ProLiant BL460
The Lenovo ThinkSystem SN550 is part of the Lenovo Flex System orchestration platform, which claims the ability to run applications with up to 80% better density than standard rack deployments. These flexible blade servers are optimized for cloud, server virtualization, database and virtual desktop infrastructure. They beat out much of the competition in scalability, supporting up to 28 cores per CPU.
Lenovo acquired the ThinkSystem line from IBM in 2015. As such, the underlying guts of the system are sound. For those businesses with an IBM/Lenovo infrastructure already in place, the SN550 is likely to be a good choice.
Read our in-depth coverage of the Lenovo ThinkSystem SN550
Cisco UCS B200 M5
Cisco UCS B200 M5 blades are rated ahead of the HPE BladeSystem and Dell PowerEdge blades by users, and only slightly behind the higher-end HPE Synergy. UCS blades are a smart choice for those already operating on Cisco. They are also a good option for those focused on networking and compute rather than storage.
But for those heavily invested in HPE or Dell systems, the storage options offered by Cisco may not be as comprehensive. If you are going to go with Cisco UCS blades, then go all-in on the Unified Computing System vision, as it ties together all aspects of compute, networking and storage throughout the enterprise.
Read our in-depth coverage of the Cisco UCS B200 M5
Huawei FusionServer E9000
The Huawei FusionServer E9000 converged-architecture blade server is a four-socket blade that goes up against the Dell PowerEdge MX480C and HPE Synergy 660 in that upper-end category. Its infrastructure was designed to converge computing, storage, networking and management.
It easily beats both of its top competitors on price but does lag behind on the number of memory and processor cores. Yet for those not already heavily invested in Dell or HPE, this may be one to try.
Read our in-depth coverage of the Huawei FusionServer E9000
Fujitsu Primergy BX400 S1
The Fujitsu Primergy BX400 S1 blade could be considered a data center on wheels. It goes up against the HPE Proliant BL460c with the c3000 enclosure. Pricing is similar and HPE comes out ahead on user ratings. But the Primergy blade wins on the number of cores.
This dual-socket server blade is probably best for SMBs and smaller projects in mid-sized organizations addressing web infrastructure workloads, and HPC. However, users may not see the same level of high support as offered by U.S.-based companies.
Read our in-depth coverage of the Fujitsu Primergy BX400 S1
NEC Express5800 Blade Enclosure M
The NEC Express5800 Blade Enclosure M is one of the cheapest blade servers on the list, making it a good option for SMBs looking to reduce their computing footprint or upgrade an aging server infrastructure. It was also built for easy setup and configuration connectivity through its management module, again making it a good fit for organizations with less server expertise.
Its managing, networking and storage options are not enough to support the needs of large enterprises. For those already invested in IBM/Lenovo, HPE, Dell, or Cisco architectures, adding another vendor may be an unwanted risk.
Read our in-depth coverage of the NEC Express5800 Blade Enclosure M
Dell PowerEdge MX840C
User reviews place the Dell PowerEdge MX840C a little behind blades from HPE and Cisco. However, this blade sits at the higher end of the Dell blade portfolio, and its only real rival in this guide is the HPE Synergy blade.
The Dell PowerEdge MX840C tops the charts on memory, compute, and storage courtesy of its double-width form factor. It may be pricier than most of the other blades, but it is a lot cheaper than the Synergy model. Ideal for high-performance applications where high density is not a major requirement.
Read our in-depth coverage of the Dell PowerEdge MX840C
HPE Synergy 660
The HPE Synergy 660 is one of three blades in the guide that can host four Xeon Scalable processors. It’s the king in terms of enterprise features and price. No other blade in this guide comes even close to offering the price tag relative to the capabilities of this high-performance model.
With the efficiency and control of the HPE Synergy 660, teams can quickly deploy IT resources for any workload through a single interface. Its composable infrastructure provides a fluid pool of physical and virtual resources that can be configured to meet the needs of virtually any application. This is our top pick for demanding workloads in an enterprise environment.
Read our in-depth coverage of the HPE Synergy 660
If you are looking for an inexpensive way to adopt four-socket Xeon blades, the SuperMicro SuperBlade fits the bill. It offers them in a much smaller form factor than its Dell and HPE competitors.
However, the SuperBlade lacks the management and enterprise-class capabilities offered by those giants. Budget-constrained organizations and SMBs should put this on their shortlist for compute-intensive blades at a reasonable cost.
Read our in-depth coverage of the SuperMicro SuperBlade
While most blade server options are scalable with the purchase of more blades and modules, there is a high probability of vendor lock-in. While the general guts of blades are similar to each other, most modules are designed to function specifically with chassis from the same vendor. So if you decide to switch vendors for part of your system, be prepared to overhaul much of the technology in your data center.