Big Data, Big Demand: Navigating the Cloud Storage Landscape

By Nicole Hemsoth

February 17, 2011

A rapid-fire search for the terms “big data” and “cloud storage” will reveal no shortage of options for users in need of a secure place to store and quickly access critical information. As the data deluge continues to slide a never-ending swell into already overstuffed datacenters, an increasing number of organizations are looking to the cloud to handle their massive demands—and for some, their processing needs as well.

A number of enterprise users are completely reliant on massive data wells to drive their businesses, and like high performance computing users, they have unique concerns when it comes to storage. However, given that the space is rather new for a an ever-diversifying breed of applications, storage concerns are overlooked in many conversations about big data.

To clear the clouds that muddle the big picture view of the storage landscape, we asked a number of big data storage experts about how they should evaluate cloud storage options and what the future of cloud storage looks like for both those with massive volumes of data to contend with as well as high performance computing users.

This week technology leaders from Panasas, EMC, NetApp, Cirtas, TwinStrata, Cleversafe, Virident, and Infineta Systems weighed in on a particular element involving cloud storage–from the nature of public clouds to scalability and cost concerns, to larger trends that are affecting the decision-making process.

Big Data and the Public Cloud Storage Problem

Garth Gibson, founder at CTO of Panasas provided a very directly-stated perspective on cloud storage, especially on public cloud storage. Gibson claims that the fundamental challenge with public cloud resources is that despite their focus on computation, the important element of storage is considered only as an afterthought.

“An opposing perspective is what is necessary for data-intensive HPC workloads. Here, big data is the true asset and computation is just part of the infrastructure. So instead of looking to potential big data applications to justify buying into the latest market hype on utility cloud computing in a far off state or country, the aggressive innovator focuses on raising top-line revenue and will spec the infrastructure (compute clusters) to the needs of the asset (big data).  Some leasing company nearby, working with a colo facility and an integrator pouncing on all things posing as cloud software, will be more than happy to build and operate the private cloud appropriate for your Big Data.  And at a reasonable price to boot.

Let’s get it straight. Understand your big data — what is it, where is it, what is needed to extract value out of it. And then build around it the private cloud best suited for it.”

Gibson remarked on the paradoxical problem with storing large amounts of data, stating that big data is “difficult and time-consuming to create, awkward to manage, expensive and slow to move, critical to gaining a competitive edge, and far from technically mature.”

This issue has been echoed from a number of quarters; the data is the central advantage for many businesses, yet it is also incredibly burdensome. Without solutions that are mature this creates a giant problem for many enterprise users. Furthermore, as Gibson asks, given the value of this data:

“Why would anyone shoehorn something that important into a public cloud—some arbitrary compute cluster at the other side of who knows where just to shave a few bucks off the apparent cost?” If high network bandwidth costs don’t kill the apparent advantage, the risks and weaknesses of privacy, efficiency, programmability, interactivity and vendor lock-in certainly do.”

While saving money is certainly a key to cloud computing decisions of any variety, Gibson’s point about the challenges for such cloud storage options are echoed by a number of other experts.

Adaptability and Cloud Storage

Along with the number of problems with cloud storage Gibson noted above, the issue of adaptability frequently emerges.

The process of choosing among vendors is already difficult given the multitude of different (yet highly similar in many ways) offerings and it can be difficult to determine how a chosen cloud storage solution is adaptable as needs change.

EMC recently acquired Isilon for its scale-out NAS offering for big data and file-based cloud opportunities. Director of Product Management for Isilon, Nick Kirsch, who was one of the engineers behind the first generation of their OneFS operating system weighed in on how high performance computing in particular can benefit from cloud storage and the scalability it provides. In Kirsch’s view,

“Given the often heterogeneous nature of HPC environments, maintaining flexibility and choice in your IT infrastructure is critical, especially in storage – where traditional systems often lock users into technology that will eventually require a fork-lift upgrade. While each form of cloud storage offers a solution to this problem, users need to make sure their chosen strategy can be adapted over time in an on-demand fashion as their needs change, which in an era of big data, they surely will.”

Whether users choose hosted cloud storage, an internal, private cloud, or a combination of both, the best practice is to ensure their chosen solution leverages on-demand storage resources that interoperate with server and networking technology – and vice versa – to enable greater flexibility and efficiency. To meet the needs of HPC and big data, storage must scale out without requiring more personnel and insulate users against unpredictable changes.”

Kirsch offered some advice for high performance computing users and those with big data demands, noting that it is indeed difficult to choose solutions that will be flexible. He suggests that in conversations with vendors about your cloud strategy, “keep them honest in offering products that can easily interoperate and scale in public, private or hybrid cloud environments. What is true today may not be true tomorrow, so freedom of choice is not only critical in any cloud storage strategy, but will, I believe, be a defining characteristic of success going forward.

New Generation of Applications Shaping the Cloud Storage Landscape

NetApp’s “Cloud Czar” Val Bercovici sees how a new generation of HPC and big data-driven applications are driving a more diverse cloud storage landscape, making scalability, flexibility and efficiency more important than ever. He notes that “Customers are increasingly facing a challenge in how to scale particularly in the backend with big data demands fueled by the mobility of data. To cope with these challenges, many customers are making the transition to flexible and efficient shared IT infrastructures—the foundation for cloud computing.

In Bercovici’s view, high-performance computing users actually have an excellent pedigree for Big Data applications since the storage layers for both share many common attributes. Big data platforms and applications predominantly depend on large sequential I/O transfers, not unlike rendering, seismic analysis or chip simulation.  Consequently high bandwidth storage systems found in high-performance computing environments are well suited for big data applications.

Security and Cost Considerations of Cloud Storage 

According to Cleversafe CEO, Chris Gladwin, the keywords for cloud storage are security and scalability. In Gladwin’s view, to receive the benefits of distributed collaboration, “cloud storage must be able to intrinsically secure the data even when it traverses the public internet.”

Gladwin notes that cloud storage has to be massively scalable—in the petabyte range and beyond but that there are considerations that users with big data must make that can influence cost. In his view, “cloud storage for big data must protect data without using replication since replication is simply too costly in petabyte scale. All this adds up to an architecture where data is securely virtualized and stored not as actual data, but instead stored on servers around the world.”

Trends on the Horizon for Big Data and Cloud Storage

Shirish Jamthe, Global Systems Director at Virident Systems sees changes in the entire landscape driving alterations in storage options as data growth increases. He notes that along with this mounting of data volumes comes a shift in terms of how applications interact with data, noting:

“In contrast to traditional data‐intensive applications such as database analytics, which have evolved over the years into disk‐friendly access patterns (medium‐to‐large sized blocks, sequential scans, seek penalty‐aware data structures, etc.), a new breed of data-intensive applications are emerging, which impose different, more challenging performance requirements on data storage devices. Examples of such applications include search, messaging, ad hoc data analytics, and social graph traversals – these applications tend to typically interact with unstructured data, performing a large number of small granularity accesses to random locations in the data collection. Such access patterns are markedly different from the carefully coordinated large-granularity accesses that characterize the traditional structured data applications.

Both of these trends–growth of data volumes, and the growing random access nature of applications outstripping advances in mechanical hard‐drive technologies–are exposing shortcomings in traditional disk drive‐based storage infrastructures, which are unable to keep up with the growing performance demands. Even currently, the most performance-intensive applications require fairly complex storage deployments, involving hundreds to thousands of disk spindles, and use of techniques such as multi-way striping and “short stroking” to meet the bandwidth and IOPS (I/O Operations Per Second) requirements, respectively. The cost, footprint, power requirements, and operational complexity of such deployments are fast making them untenable for the wide variety of situations that demand such performance levels.”

In Jamthe’s view the traditional SAN & NAS solutions have worked very well for companies for several years but have now started to show limitations against this explosive growth. The gap between DRAM capacity and disk storage has increased reducing the efficiency of DRAM cache. A new class of memory Storage-Class Memories (SCM) offers what he terms, “an interesting blending of the characteristics of both DRAM and disk. Storage built using these devices offer performance and random-access capabilities of DRAM, and the persistence and capacity of disk drives. NAND Flash-based Solid State Drive (SSD) is the first instance of SCM that is going main stream.” Of these solutions, Jamthe notes,

“A large number of companies, ranging from non-volatile memory manufacturers to server OEMs are offering NAND Flash-based SSDs in various form factors. All of these products are riding the Moore’s Law influenced scaling of NAND Flash capacities (with accompanying reductions in cost/GB), which is making possible disk drive‐like capacities for a moderate cost multiple. When one takes into consideration the hundred to thousand-fold improvements in random access performance possible with NAND Flash devices and their sharply dropping cost, it appears inevitable that SSDs will replace traditional disk drives, used as the performance storage tier, in majority of data-intensive applications in a few years.”

What’s Missing from Cloud Storage Conversations?

Haseeb Budhani, VP of Products at Infineta Systems notes that as organizations of all sizes (from research labs to enterprises) beginning to experiment with the likes of Hadoop, Cassandra, and other custom “distributed” or scatter-gather applications, the storage requirements to handle these large workloads are also growing steadily. He says that “storage growth is leading both enterprise, who have traditionally purchased from vendors such as EMC and IBM, to explore cheaper and more scalable storage alternatives. Enter cloud storage.”

In Budhani’s view the industry is focusing on the quickly-growing big data movement, “but the storage aspect of this movement doesn’t seem to have the industry’s full attention today. The storage problem is not simply I/O related (i.e. the speed at which increasingly fast computations need to access data), but also has to do with how much data is being stored, and where.” He goes on to offer the following opinon about what is missing from big data conversations:

With all this “big data” being generated and accessed at very high speeds, new concerns are quickly taking the spotlight. Some of these have to do with the fact that:

a) Access to this data from the customer’s primary business location needs to be reliable and fast;

b) When a customer’s data is “in the cloud,” it may technically reside across multiple locations; and

c) Depending on where the computation is taking place in the cloud, fast access to all of the customer’s data is critical.

The first issue above has been effectively addressed by the likes of Cirtas and StorSimple, and it remains to be seen whether businesses will continue to locally access data in the future. If all applications and requisite data reside in the cloud, there may not be any need for a cloud storage gateway.

The other two issues have been mostly left unaddressed to date. With data being housed across locations, it is essential that the WAN connectivity between these locations be extremely fast. Depending on the geographical location of a data center, it may not be trivial to provision additional bandwidth in short order. What’s more, in some cases, it is not the lack of bandwidth that is the concern, but the latency impact of using transport protocols such as TCP across sites that become a bottleneck.

The Final Word on Big Data and Cloud Storage

The jury is still out for high-performance computing users with applications that would suffer from cloud storage use but for those with massive data analytics challenges, there are benefits that can be realized in the arenas of scalability. One of the first questions that decision makers need to consider is what the tangible core benefit is to choosing cloud storage.

Josh Goldstein, who runs product management for Cirtas notes that there are relatively simple questions behind whether or not cloud storage for data-driven applications is a good route to pursue. Will the data be accessed frequently or does it just need to be kept? Secondly, if the data will be accessed, are the servers doing the processing in the cloud?

Goldstein suggests that “when data is infrequently accessed, leveraging public cloud storage via a cloud storage controller (known as a hybrid cloud model) offers attractive cost, data protection, and simplification benefits. Infrequent access is an important criteria here because moving “big data” back and forth to the cloud in short time periods is impractical.

He goes on to advise that when access to data is paramount, the decision boils down to where the processing will take place. “Once data is in the cloud, operating on it using cloud compute farms is very compelling, especially when the compute needs are elastic.  If the intention is to use local compute resources, a private cloud storage model makes more sense as it avoids the large WAN transfers that might slow data processing.”

According to John Bates, co-founder and CTO at TwinStrata, Inc, there are clear benefits to using cloud versus physical storage for HPC users or those with big data problems to solve. Bates notes that for these challenges, “cloud storage can solve some of the problems associated with big data, particularly in the areas of resource planning and infrastructure growth costs. Cloud storage offers massive and automatic scalability, without requiring heavy capital expenditures on fixed storage systems that may reach capacity too fast.

Bates also feels that storing data in the cloud also enables more agility, leading to new service opportunities for big data analysis, such as in the field of business intelligence.

This concludes our survey of leaders from storage companies but we’d like to hear from you. If you’re considering the possibility of cloud storage for high-performance computing or big data needs, what are some of the essential questions that never get answered?

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industy updates delivered to you every week!

Weekly Twitter Roundup (Feb. 23, 2017)

February 23, 2017

Here at HPCwire, we aim to keep the HPC community apprised of the most relevant and interesting news items that get tweeted throughout the week. Read more…

By Thomas Ayres

HPE Server Shows Low Latency on STAC-N1 Test

February 22, 2017

The performance of trade and match servers can be a critical differentiator for financial trading houses. Read more…

By John Russell

HPC Financial Update (Feb. 2017)

February 22, 2017

In this recurring feature, we’ll provide you with financial highlights from companies in the HPC industry. Check back in regularly for an updated list with the most pertinent fiscal information. Read more…

By Thomas Ayres

Rethinking HPC Platforms for ‘Second Gen’ Applications

February 22, 2017

Just what constitutes HPC and how best to support it is a keen topic currently. Read more…

By John Russell

HPE Extreme Performance Solutions

O&G Companies Create Value with High Performance Remote Visualization

Today’s oil and gas (O&G) companies are striving to process datasets that have become not only tremendously large, but extremely complex. And the larger that data becomes, the harder it is to move and analyze it – particularly with a workforce that could be distributed between drilling sites, offshore rigs, and remote offices. Read more…

HPC Technique Propels Deep Learning at Scale

February 21, 2017

Researchers from Baidu’s Silicon Valley AI Lab (SVAIL) have adapted a well-known HPC communication technique to boost the speed and scale of their neural network training and now they are sharing their implementation with the larger deep learning community. Read more…

By Tiffany Trader

IDC: Will the Real Exascale Race Please Stand Up?

February 21, 2017

So the exascale race is on. And lots of organizations are in the pack. Government announcements from the US, China, India, Japan, and the EU indicate that they are working hard to make it happen – some sooner, some later. Read more…

By Bob Sorensen, IDC

ExxonMobil, NCSA, Cray Scale Reservoir Simulation to 700,000+ Processors

February 17, 2017

In a scaling breakthrough for oil and gas discovery, ExxonMobil geoscientists report they have harnessed the power of 717,000 processors – the equivalent of 22,000 32-processor computers – to run complex oil and gas reservoir simulation models. Read more…

By Doug Black

TSUBAME3.0 Points to Future HPE Pascal-NVLink-OPA Server

February 17, 2017

Since our initial coverage of the TSUBAME3.0 supercomputer yesterday, more details have come to light on this innovative project. Of particular interest is a new board design for NVLink-equipped Pascal P100 GPUs that will create another entrant to the space currently occupied by Nvidia's DGX-1 system, IBM's "Minsky" platform and the Supermicro SuperServer (1028GQ-TXR). Read more…

By Tiffany Trader

HPC Technique Propels Deep Learning at Scale

February 21, 2017

Researchers from Baidu’s Silicon Valley AI Lab (SVAIL) have adapted a well-known HPC communication technique to boost the speed and scale of their neural network training and now they are sharing their implementation with the larger deep learning community. Read more…

By Tiffany Trader

IDC: Will the Real Exascale Race Please Stand Up?

February 21, 2017

So the exascale race is on. And lots of organizations are in the pack. Government announcements from the US, China, India, Japan, and the EU indicate that they are working hard to make it happen – some sooner, some later. Read more…

By Bob Sorensen, IDC

TSUBAME3.0 Points to Future HPE Pascal-NVLink-OPA Server

February 17, 2017

Since our initial coverage of the TSUBAME3.0 supercomputer yesterday, more details have come to light on this innovative project. Of particular interest is a new board design for NVLink-equipped Pascal P100 GPUs that will create another entrant to the space currently occupied by Nvidia's DGX-1 system, IBM's "Minsky" platform and the Supermicro SuperServer (1028GQ-TXR). Read more…

By Tiffany Trader

Tokyo Tech’s TSUBAME3.0 Will Be First HPE-SGI Super

February 16, 2017

In a press event Friday afternoon local time in Japan, Tokyo Institute of Technology (Tokyo Tech) announced its plans for the TSUBAME3.0 supercomputer, which will be Japan’s “fastest AI supercomputer,” Read more…

By Tiffany Trader

Drug Developers Use Google Cloud HPC in the Fight Against ALS

February 16, 2017

Within the haystack of a lethal disease such as ALS (amyotrophic lateral sclerosis / Lou Gehrig’s Disease) there exists, somewhere, the needle that will pierce this therapy-resistant affliction. Read more…

By Doug Black

Azure Edges AWS in Linpack Benchmark Study

February 15, 2017

The “when will clouds be ready for HPC” question has ebbed and flowed for years. Read more…

By John Russell

Is Liquid Cooling Ready to Go Mainstream?

February 13, 2017

Lost in the frenzy of SC16 was a substantial rise in the number of vendors showing server oriented liquid cooling technologies. Three decades ago liquid cooling was pretty much the exclusive realm of the Cray-2 and IBM mainframe class products. That’s changing. We are now seeing an emergence of x86 class server products with exotic plumbing technology ranging from Direct-to-Chip to servers and storage completely immersed in a dielectric fluid. Read more…

By Steve Campbell

Cray Posts Best-Ever Quarter, Visibility Still Limited

February 10, 2017

On its Wednesday earnings call, Cray announced the largest revenue quarter in the company’s history and the second-highest revenue year. Read more…

By Tiffany Trader

For IBM/OpenPOWER: Success in 2017 = (Volume) Sales

January 11, 2017

To a large degree IBM and the OpenPOWER Foundation have done what they said they would – assembling a substantial and growing ecosystem and bringing Power-based products to market, all in about three years. Read more…

By John Russell

US, China Vie for Supercomputing Supremacy

November 14, 2016

The 48th edition of the TOP500 list is fresh off the presses and while there is no new number one system, as previously teased by China, there are a number of notable entrants from the US and around the world and significant trends to report on. Read more…

By Tiffany Trader

Lighting up Aurora: Behind the Scenes at the Creation of the DOE’s Upcoming 200 Petaflops Supercomputer

December 1, 2016

In April 2015, U.S. Department of Energy Undersecretary Franklin Orr announced that Intel would be the prime contractor for Aurora: Read more…

By Jan Rowell

D-Wave SC16 Update: What’s Bo Ewald Saying These Days

November 18, 2016

Tucked in a back section of the SC16 exhibit hall, quantum computing pioneer D-Wave has been talking up its new 2000-qubit processor announced in September. Forget for a moment the criticism sometimes aimed at D-Wave. This small Canadian company has sold several machines including, for example, ones to Lockheed and NASA, and has worked with Google on mapping machine learning problems to quantum computing. In July Los Alamos National Laboratory took possession of a 1000-quibit D-Wave 2X system that LANL ordered a year ago around the time of SC15. Read more…

By John Russell

Enlisting Deep Learning in the War on Cancer

December 7, 2016

Sometime in Q2 2017 the first ‘results’ of the Joint Design of Advanced Computing Solutions for Cancer (JDACS4C) will become publicly available according to Rick Stevens. He leads one of three JDACS4C pilot projects pressing deep learning (DL) into service in the War on Cancer. Read more…

By John Russell

IBM Wants to be “Red Hat” of Deep Learning

January 26, 2017

IBM today announced the addition of TensorFlow and Chainer deep learning frameworks to its PowerAI suite of deep learning tools, which already includes popular offerings such as Caffe, Theano, and Torch. Read more…

By John Russell

HPC Startup Advances Auto-Parallelization’s Promise

January 23, 2017

The shift from single core to multicore hardware has made finding parallelism in codes more important than ever, but that hasn’t made the task of parallel programming any easier. Read more…

By Tiffany Trader

CPU Benchmarking: Haswell Versus POWER8

June 2, 2015

With OpenPOWER activity ramping up and IBM’s prominent role in the upcoming DOE machines Summit and Sierra, it’s a good time to look at how the IBM POWER CPU stacks up against the x86 Xeon Haswell CPU from Intel. Read more…

By Tiffany Trader

Leading Solution Providers

Tokyo Tech’s TSUBAME3.0 Will Be First HPE-SGI Super

February 16, 2017

In a press event Friday afternoon local time in Japan, Tokyo Institute of Technology (Tokyo Tech) announced its plans for the TSUBAME3.0 supercomputer, which will be Japan’s “fastest AI supercomputer,” Read more…

By Tiffany Trader

Nvidia Sees Bright Future for AI Supercomputing

November 23, 2016

Graphics chipmaker Nvidia made a strong showing at SC16 in Salt Lake City last week. Read more…

By Tiffany Trader

BioTeam’s Berman Charts 2017 HPC Trends in Life Sciences

January 4, 2017

Twenty years ago high performance computing was nearly absent from life sciences. Today it’s used throughout life sciences and biomedical research. Genomics and the data deluge from modern lab instruments are the main drivers, but so is the longer-term desire to perform predictive simulation in support of Precision Medicine (PM). There’s even a specialized life sciences supercomputer, ‘Anton’ from D.E. Shaw Research, and the Pittsburgh Supercomputing Center is standing up its second Anton 2 and actively soliciting project proposals. There’s a lot going on. Read more…

By John Russell

IDG to Be Bought by Chinese Investors; IDC to Spin Out HPC Group

January 19, 2017

US-based publishing and investment firm International Data Group, Inc. (IDG) will be acquired by a pair of Chinese investors, China Oceanwide Holdings Group Co., Ltd. Read more…

By Tiffany Trader

Dell Knights Landing Machine Sets New STAC Records

November 2, 2016

The Securities Technology Analysis Center, commonly known as STAC, has released a new report characterizing the performance of the Knight Landing-based Dell PowerEdge C6320p server on the STAC-A2 benchmarking suite, widely used by the financial services industry to test and evaluate computing platforms. The Dell machine has set new records for both the baseline Greeks benchmark and the large Greeks benchmark. Read more…

By Tiffany Trader

TSUBAME3.0 Points to Future HPE Pascal-NVLink-OPA Server

February 17, 2017

Since our initial coverage of the TSUBAME3.0 supercomputer yesterday, more details have come to light on this innovative project. Of particular interest is a new board design for NVLink-equipped Pascal P100 GPUs that will create another entrant to the space currently occupied by Nvidia's DGX-1 system, IBM's "Minsky" platform and the Supermicro SuperServer (1028GQ-TXR). Read more…

By Tiffany Trader

What Knights Landing Is Not

June 18, 2016

As we get ready to launch the newest member of the Intel Xeon Phi family, code named Knights Landing, it is natural that there be some questions and potentially some confusion. Read more…

By James Reinders, Intel

Is Liquid Cooling Ready to Go Mainstream?

February 13, 2017

Lost in the frenzy of SC16 was a substantial rise in the number of vendors showing server oriented liquid cooling technologies. Three decades ago liquid cooling was pretty much the exclusive realm of the Cray-2 and IBM mainframe class products. That’s changing. We are now seeing an emergence of x86 class server products with exotic plumbing technology ranging from Direct-to-Chip to servers and storage completely immersed in a dielectric fluid. Read more…

By Steve Campbell

  • arrow
  • Click Here for More Headlines
  • arrow
Share This