The first two things that people usually think of when they consider SSDs are that they are both much faster and more expensive than HDDs at the same level of drive capacity. Their high price tag makes SSDs currently under-utilized in the market even though for some applications you’d be saving money by using them. Let’s briefly look at a basic example.

Let’s assume we’re currently running the hardware configuration below:

• Configuration 1
• 10x 2U servers
• 24 SAS drives
• $80,000 total cost ($36,000 drives + $40,000 servers + $4,000 rack space)

In the above example, let’s assume we decided to go with SAS drives because our applications demand high IOPS but not necessarily much storage space (i.e. we’re currently under-utilizing the storage capacity). In such a scenario, it is much more cost-effective to have used SSDs rather than SAS drives in designing a configuration. Here is how the improved configuration would look:

• Configuration 2
• 4x 1U servers
• 8 SSD drives
• $30,800 total cost ($5,000 drives + $25,000 servers + $800 rack space)

Configuration 2, which uses SSDs instead of HDDs, saves you 62% in purchasing costs over Configuration 1 (it will also save you money in operating costs because SSDs use less energy than spinning drives and there are also less total servers to power). This simple example illustrates some of the possibilities for hardware cost and energy consumption savings that SSDs promise, but deciding on whether to go with SSDs or HDDs is all about choosing the right tool for the job. Let’s look at this challenge in more detail.

There are three key metrics (aside from price) to consider when determining whether SSDs are the smart option for your computing system. These are storage, IOPS, and workload requirements. Let’s look at storage capacity first.

Storage Capacity

It’s undeniable that if your applications require large quantities of storage space, SAS and SATA drives look much more appealing than SSDs to the pocketbook. As Chart A demonstrates, for the same GB capacity, an SSD drive is much pricier than a SAS drive (and the SSDs don’t even come close to matching the capacity levels of SATA drives).

Chart A

So, if I/O performance is not important and you require merely storage capacity, SATA or SAS drives are usually the best options. But what if your applications call for both many gigabytes and high I/O?

You can attain a solution for this situation by designing a hybrid system, a server that uses both SSDs and HDDs. Although this requires more in-depth knowledge of your applications’ resource utilization to implement, such a solution can make your servers more efficient and save you money. By assigning some tasks (such as handling swap files and transaction logs) to SSDs and other storage tasks to HDDs, you can reduce the I/O bottleneck and at the same time allow for large quantities of data to be stored on the system.

There is another type of solution with hybrid drives, which works exceptionally well in storage systems where the need for drive read (but not write) performance is extremely high. This solution is based on LSI’s CacheCade technology that uses SSDs as RAID controller cache. Read more about such an implementation here.

Once you’ve considered your storage requirements, the next factor to take into account is performance.

IOPS

IOPS stands for Input/Output Operations Per Second (IOPS), and it is a measure of drive performance. For HDDs, IOPS is largely determined by the quickness with which the rotating disk and the read/write head operate. Since SAS drives spin at speeds much higher than SATA drives (15,000 RPM vs. 7,200 RPM, respectively), you can frequently rule out SATA drives as an option if your applications require high IOPS.

SSDs can attain comparable or better IOPS levels as SAS drives for a much lower cost. So, when it comes to IOPS, the typical mantra that “SSDs are too expensive” is actually reversed. SSDs only seem expensive when comparing price per gigabytes (see Chart A); if you compare price per IOPS, SAS drives are the more costly option (see Chart B).

Chart B

Chart B also shows FIO in addition to SSD and SAS drives. FIO stands for Fusion-io, a vendor of a unique and high-end type of I/O solution. FIO uses solid-state flash memory connected via a PCIe slot on the motherboard to achieve perhaps the highest levels of IOPS possible in a system (over 1,000,000 IOPS, in some cases). Such a solution is very pricey and is only recommended in extreme cases where the application calls for the highest performance available regardless of the cost (e.g. computational finance).

But IOPS is still only one part of the I/O equation. You still have to consider the workload requirements of your applications.

Sequential vs. Random I/O

I/O operations can be classified as one of two types: sequential or random. In HDDs, the most time consuming part of the I/O process is when the disk head seeks the disk cylinder to access the correct data. If it needs to access data that is scattered in many different places on the disk (i.e. random I/O workload), then HDD performance will be significantly slower. On the other hand, if the data to be read is located in the same general area of the disk cylinder (and the disk is not terribly fragmented), then HDDs can tackle such sequential I/O workloads rather well.

Applications that use sequential I/O operations include backup, archiving, and streaming video. Two common examples of random I/O are database servers and Microsoft Exchange servers.

Since SSDs don’t have spinning parts, they perform random I/O operations better than hard disk drives. Chart C approximately demonstrates this workload to technology alignment.

Chart C

Of course the devil is in the details, and you should  first determine specifically what type of workload your applications will demand. As a general rule of thumb, SATA drives are sufficient for situations which primarily require sequential I/O, SSDs should be used when faster seek time is at a premium (random I/O), and SAS drives are a good solution when you expect a mixed workload.

Conclusion: Where SSDs excel

The best situations where you could use solid-state drives are ones where:

1. The size of your data set is small,
2. You require high IOPS, and
3. Your I/O workload is mostly random operations.

There are other important factors to consider when deciding whether SSDs or HDDs are the right tool for the job. For instance, most types of enterprise-level SSDs have write longevity constraints over the course of their usable life. In addition, if you decide to reduce the number of servers you deploy by utilizing SSDs, you should make sure that scaling down your hardware does not create other bottlenecks in processing power, network I/O, or RAM.

If you have any other advice for those trying to decide between SSDs and HDDs when designing computing systems, please post in the comments below. Thanks for reading!

* This article (and the data mentioned) is based on an excellent presentation delivered by Douglas Bone on April 19, 2011.

If you haven’t heard the news yet, the next generation Sandy Bridge Intel Xeon E5 processors have been released for the general public. We’ve announced our new ‘Cosmos’ Server line, boasting the Xeon E5-2600 family of processors.

Learn more about them on our website and contact us to put the power of the ‘Cosmos’ at the core of your computing today!

The promotion includes this hardware:

• Case: Supermicro 813MT-350CB
• Motherboard: X9SCL-F
• Processor: Intel Xeon E3-1230 3.2 GHz
• Power Supply: 350W High-Efficiency

You also get your choice of 8GB or 16GB memory and up to 8 TB of storage space. Our promotion includes access to phenomenally-priced Toshiba enterprise hard drives.

It’s not social shopping without your engagement! Don’t be anti-social… share with others and claim your 20% discount today!

SATA and SAS are storage interface and bus types designed to aid in the movement of data from one place to another. Think of SAS and SATA as different kinds of computer interfaces, such as PCI Express, but there are actually multiple components that make up the overall SAS architecture.

• Initiators. The initiator is the SAS controller to which SAS expanders or targets can be connected.
• Expanders. Expanders sit between initiators and targets, but can also connect to other expanders, as you can see in Figure A. Expanders are sort of like network switches in that they can direct traffic and they provide the ability to scale the SAS architecture beyond single port limits.
• Targets. A target is either a SAS drive or a SATA disk. SATA disks can be connected to SAS expanders and initiators, but do not perform quite as well as SAS disks.

Each drive has its own host interface (SAS or SATA) which are connected to an initiator or an expander. The initiator, functioning as the system controller, communicates with the SSD using the appropriate commands as dictated by the host interface (again, SAS or SATA.)

In other words, the SSD functions no differently than a hard-disk in terms of how it the interfaces (SAS or SATA) work. It simply utilizes a different method of storing the data.

Now… let’s introduce iSCSI

The above image includes an iSCSI initiator in the host server:

It’s either hardware- or software-based and controls communications with the Ethernet-based storage array. By virtue of the fact that the system is using Ethernet for storage traffic, speeds of both 1 Gbps and 10 Gbps are supported. In an iSCSI storage network, SCSI commands originate on the host, at which point, they are encapsulated inside TCP/IP packets. These TCP/IP packets then traverse the Ethernet network just like any other network traffic. At the storage array side of the equation, these packets are received and the SCSI commands are extracted from the TCP/IP packet and passed on to the storage device for execution. Data that is transmitted back to the host undergoes this same encapsulation process.

Basically, iSCSI adds some extra processor load because of the encapsulation process which takes time. We can think of iSCSI as a “storage transportation method.” In that sense, they resemble SATA and SAS in dealing with how the data is moved around (vs. SSD which deals with how the data is stored.) While this adds resource consumption, the author notes that it also provides a good method to decouple storage from the host, allowing for the share of storage among multiple hosts, as shown below:

However, you might be wondering exactly how does iSCSI compare to SATA and SAS. The author begins to address this.

iSCSI is the storage transport but, at present, there is no such thing as an “iSCSI disk.” iSCSI arrays use SAS and SATA disks (which can also be SSDs) for storage, but the data is transported to hosts using iSCSI.

While SAS and SATA or networking interfaces, iSCSI is a networking protocol that can run across the infrastructure. In that sense, iSCSI is complementary, not contradictory, to SAS and SATA.

Check out the original article over at TechRepublic.

Die shot of 'Aubrey Isle,' the silicon chip included in the Intel MIC 'Knights Ferry' development platform

Intel “MIC” (for Many Integrated Core) co-processor architecture is a leading initiative in Intel’s move into the realm of HPC and parallel computing. At Supercomputing 2011, Intel had an opportunity to show off the advantages of Intel MIC for a variety of computing tasks, including weather modeling, tomography, protein folding and advanced materials simulation. Intel beamed with joy unveiling their 50 core “Knights Ferry” prototype last year.

This year we are expected to see the next product in this line: the much-anticipated commercial “Knights Corner.” Early presentations of “Knights Corner” showed that the Intel architecture is capable of delivering over 1 TFLOPs of double precision floating point performance, the highest performance level achieved for a single processing chip. With this, Intel will finally have an answer to NVIDIA’s Tesla line and AMD’s Fusion APUs.

Intel MIC is also part of a larger strategy which has become clearer in recent weeks. With the acquisition of the Infiniband technology line from QLogic back in January of 2012 and the announcement of the Crystal Forest platform earlier this week, it is obvious that Intel hopes 2012 will be the year it starts to establish a solid position in the high performance computing space.

Despite the fact that we are on the cusp of the Romley / Xeon E5 release, this almost overshadows that launch in a way because it is indicative of a serious expansion for Intel into the HPC, networking, heterogeneous computing, and data center industries. Intel has even claimed that MIC outperforms GPUs when it comes to parallel computing, which means that NVIDIA in particular, as the leading provider of GPU computing solutions, may have some serious competition to contend with in the near future. And, with dedicated chipsets geared towards increasing data center network speeds without adversely affecting performance or security, it’s not hard to see how Intel’s strategy is intended to play out.

When you couple all of this with the fact that, traditionally, Intel has risen to be the dominant force in virtually every market it has entered, the future of enterprise IT has some very interesting possibilities.

Other comments made by Raj Hazra (either in the video or in the original Intel press release):

• “Knights Corner,” will be manufactured using Intel’s latest 3-D Tri-Gate 22nm transistor process. Featuring over 50 cores, Intel MIC will offer both high performance from an architecture specifically designed to process highly parallel workloads and compatibility with existing x86 programming model and tools.
• The “Knights Corner” co-processor is very unique as, unlike traditional accelerators, it is fully accessible and programmable. In this way it resembles a fully functional HPC compute node, visible to applications as though it was a computer that runs its own operating system independent of the host OS.
• A major benefit of the Intel MIC architecture is the ability to run existing applications without the need to port the code to a new programming environment. This will allow scientists to use both CPU and co-processor performance simultaneously with existing x86 based applications, dramatically saving time, cost and resources that would otherwise be needed to rewrite them to alternative proprietary languages.

HPC @ Intel 2011 – video with more info on Intel and MIC (released Summer 2011 outlining Intel’s future development in the high performance computing industry)

The promotion includes this hardware:

• Case: Supermicro 813MT-350CB
• Motherboard: X9SCL-F
• Processor: Intel Xeon E3-1230 3.2 GHz
• Power Supply: 350W High-Efficiency

You also get your choice of 8GB or 16GB memory and up to 8 TB of storage space. Our promotion includes access to phenomenally-priced Toshiba enterprise hard drives.

Remember, though, in order to get the discount we need enough people to participate! That’s where you come in! Help us spread the word by:

• Sharing on social networks like Twitter, Facebook, Google+, and LinkedIn
• Mentioning it in a relevant discussion forum
• Sending an email or phone call to someone you think could benefit from these tremendous savings
• Any other way you can think of!

It’s not social shopping without your engagement! Don’t be anti-social… share with others and claim your 20% discount today!

In the original Intel press release, the company announced that:

“Customer acceptance of the Intel Xeon E5 processor has exceeded our expectations and is driving the fastest debut on the TOP500 list of any processor in Intel’s history,” said Hazra. “Collecting, analyzing and sharing large amounts of information is critical to today’s science activities and requires new levels of processor performance and technologies designed precisely for this purpose.”

With their increasing movement into the HPC market, the release of the E5 processor series to the general public (they were shipped to a small number of cloud and HPC customers in September 2011) will help solidify Intel’s position in a specialized field of computing where their previous products have typically not done as well. At the time of Intel’s original press release, Xeon E5 processors were already powering 10 systems on the TOP500 supercomputers list, with over 20,000 of these processors in operation delivering a cumulative peak performance of over 3.4 petaflops. And, it is expected that the Xeon E5 processor family will power several other supercomputers in the future. In other words, this looks to be the beginning of something big for Intel in the realm of HPC.

A recent article from Computerworld has some good highlights on upcoming developments for the Xeon E5 launch:

Intel is actually cutting its server market into eight pieces with the Xeon E5 launch. That’s Itanium 9300s and Xeon 7500s and E7s at the high-end (and eventually the “Sandy Bridge-EX” E8s). That’s two segments of the market that share chipsets and memory cards, but that have different motherboards and sockets. At least until Intel finally delivers, as it is rumored to be in the works, the long-promised common Xeon-Itanium socket. That could happen with the E8s, but it is far more likely to happen with the “Ivy Bridge-EX” Xeon E9s years hence. At the low-end, there’s the single-socket Xeon E3 and Atom processors, depending on how wimpy or brawny your workload is. That’s four addressable server segments in total.

The Xeon E5s will also span four different server types and will cover the middle and overlap with the high and low ends. The Xeon E5-2600, as the first of the “Romley” server platforms are expected to be called, will use the “EP” variant of the Xeon E5 chip that plugs into the new “Socket R” CPU socket. This socket is not compatible with the current Xeon 5500 and 5600 processors, but has all sorts of goodies, including two QPI links between the processors, support for unregistered, registered, and load-reduced (LR) DDR3 main memory, and integrated PCI-Express 3.0 controllers on the processor. This is the chip that Intel has presumably been shipping under NDA to selected supercomputer and hyperscale data center customers since last fall. This chip is clearly aimed at two-socket Opteron 6200 machines.

For two-socket machines that don’t need all of these capabilities, Intel is expected to roll out its “Sandy Bridge-EN” chips, rumored to be called the Xeon E5-2400s. These chips will plug into the new “Socket B2″ socket and will sport only one QPI link between processors as well as fewer memory channels, fewer DIMMs per core, and fewer PCI-Express 3.0 slots. This chip is fired directly at two-socket Opteron 4200 iron.

If the rumors are right, then Intel will also ship a variant of the Sandy Bridge-EP chip that will be able to span four processor sockets in a single system image. This chip is expected to be called the Xeon E5-4600 and is obviously targeting the four-socket Opteron 6200.

Despite all of the excitement around the Xeon E5 series, this is only the most upcoming of Intel’s many launches this year. Check back next week when we will take a look at another one of Intel’s big plans: MIC.

But of all the areas in tech innovation, storage may be the one to set itself apart this year. Yesterday Computerworld ran a very thorough and informative piece titled “2012: The year storage becomes a celebrity,” in which they laid out some of the things to look forward to in the development of enterprise storage this year.

The Overview

While data storage has always been a necessary building block for technology, it’s rarely garnered as much attention as it has in the past two years. The reason: Corporate and retail consumers are being forced to store greater amounts of data and they need to make that data more useful — and accessible.

Enterprise storage, and storage in general, is one of those things that trends dictate will get increasingly cheaper. That doesn’t always mean that it will become more accessible, as storage needs may outpace dropping costs. However, in general storage has become more and more affordable, making a big difference to smaller and even mid-sized organizations. It’s also something that is shared across all industries almost without exception. Storage is not just an “IT” issue, but rather plays a critical role in healthcare, entertainment, finance, engineering, etc.

And ten years ago, we did not have the wide variety of options we do today…

The Rise of the SSD

One of the most closely-watched storage technologies has been the solid-state drive. When hard disk production suffered a tremendous blow from the flooding in Thailand last November, some analysts saw this as a turning point in favor of widespread SSD adoption. While the last few months have not given rise to a solid-state revolution, across-the-board price increases in hard drives made SSDs considerably more attractive as an alternative option.

It is also important to note that, while SSDs remain more expensive than HDDs, historical price trends show that their prices have fallen relatively more quickly than their hard disk counterparts.

According to new data from research firm IDC, worldwide solid-state storage industry revenue hit $5 billion in 2011, up 105% from the $2.4 billion mark in 2010. IDC expects the market will expand further in 2012 and beyond.

“2011 was a record year for the worldwide SSD market, with revenue more than doubling year over year due to strong SSD shipment growth in the enterprise and client segments,” said Jeff Janukowicz, an IDC research director. “The increasing use of flash in enterprise solutions, explosive growth of mobile client devices, and lower SSD pricing is creating a perfect storm for increased SSD shipments and revenue over our forecast.”

The Computerworld article also notes that combining these technologies together opens up possibilities that do not exist with only the traditional spinning disks. We can attest to that, as we have seen some very interesting applications done with this combination.

• Quickly making on-the-fly disk images & full copies is significantly easier than spinning drives unless they are also paired with a very efficient RAID configuration (minimum RAID 5).
• Utilizing SSDs as boot drives can greatly speed up a shared system such as a virtual private server.
• SSDs can be useful as a replication point in a MySQL dump.
• SSDs can help with caching files in a system with spinning disks, boosting performance across the entire system

The bottom line is that we now are seeing how SSDs can be used (separately and/or in conjunction with hard drives) to achieve distinct goals in a more efficient manner, rather than just using them as spinning-disk replacements.

The Future of PCIe in Storage

The other big trend is the use of PCIe technology in storage.

Dell is already selling PCIe server cards from Fusion-io, and EMC soon plans to announce the availability of its PCIe flash card initiative, dubbed Project Lightning. EMC is expected to sell its own branded flash memory cards from multiple suppliers to server makers and is likely to offer flash memory on blade servers as part of its vBlock offering.

…

Wayne Adams, chair of the Storage Networking Industry Association (SNIA), is upbeat about the PCIe server cards.

“Now, you have higher speeds and you’re looking at transferring data at line speed between a server’s microprocessor and [flash] cache,” said Adams, who attended a two-hour panel at his association’s Winter Symposium on PCIe-solid state storage standards. “System designers are working with cache and higher speed interconnects like PCIe and changing the equation of how much storage can be put in a box in order to match server computational growth. So you can end up with more efficient storage instead of 100 hard drives with data striped across them.

“It’s all about doing more for less or more in the same foot print,” he added.

To-date, most of the use of flash memory in data center computing has been for specialized applications and use in hybrid systems. The traditional barrier is price. But as the cost handicap begins to fade the benefits of faster data access and retreival, not to mention the savings from power and storage space efficiency, such technology is becoming more and more attractive.

PCIe-based SSDs and other forms of flash memory incorporate these benefits into robust solutions for enterprise, academia, and SMB. Within this line of development, the term storage class memory emerged a couple years ago to designate high-performance memory that has the capacity and economics of storage.

While many flash-based storage devices can offer compelling economics compared to DRAM, storage class memory specifically aims to provide near-RAM performance, while also attempting to minimize I/O bottleneck issues that can arise from traditional storage connection protocols such as SATA or SAS.

Two of the leading storage class memory products on the market are Virident’s FlashMax and LSI’s WarpDrive. Both offer incredible performance compared to even some of the standards of today. They are indicative of the future trends of storage.

Next on the Horizon for Storage?

Created by IBM

These are the biggest developing trends in the field of enterprise storage. But whatever specific technologies that emerge as dominant, the challenges of storage are shared by all. With an ever-growing amount of new data and analytical tools, storage evolution must be able to keep up with our data-collection, data-management, and data-crunching needs.

Storage is an essential aspect of high performance computing, a field in which big data represents a tremendous growth opportunity. To learn more, read our article “Future Trends in HPC, Part 2.”