Building a cloud computing infrastructure to serve dual purposes

We’ve been bombarded with cloud computing services terms that describe many types of services that can be offered through the cloud. For network operators, the challenge with multifaceted clouds is that they have a variety of business drivers. The catch is that if different drivers push cloud computing infrastructure in incompatible directions, the consequences could be dire for service providers’ capital and operations expenses, as well as for return on investment (ROI).

Unlike enterprise clouds, this diversity of business drivers requires service provider clouds to function as platforms for traditional OSS/BSS and internal IT, as platforms to host features and content and as platforms for cloud services for different groups of customers -- enterprises, SMBs and consumers. Virtually every operator would have a different balance of need/opportunity in each of these areas.

In addition to offering cloud services including Infrastructure as a Service (IaaS), Platform as a Service (PaaS), Software as a Service (SaaS), Database as a Service and Hosting as a Service, internally, network operators are also giant IT consumers themselves. That means they build data centers not only for OSS/BSS and traditional applications, but for hosting content and features.

Network operators have most of the same concerns as enterprises about cloud computing infrastructure efficiency, application performance and even “cloudsourcing” to third parties, which means that operators are shifting their IT investment from an internal fulfillment strategy to one hosted in the cloud. For this supply-side vision of cloud migration, the best course is often dictated by the cloud strategies supported by the major IT vendors, just as it is for enterprises.

Yet the issue of business and network transformation represents a special challenge for service providers -- restructuring their network infrastructure to meet changing revenue goals and new service targets. Five years from now, most operators will be earning the majority of their revenues from sources that were minimal or non-existent contributors five years before.

Cloud computing infrastructure’s dual role -- internal IT and customer services

The service-driven transformation of network operators is more than a network infrastructure transformation. Content and social network services, mobile features, app stores and all of the things included as elements in the future of network services are predominantly hostedrather than simply connected. That makes a provider’s IT structure as critical as its network. Hosted services and features also make the architecture that binds these IT elements into a critically important set of cohesive resources.

The cloud is important to operators because it is the abstract model that nearly all providers are selecting for their own internal IT needs, which include hosting features and content, as well as supporting their back-offices or OSS/BSS systems. If service providers were to create a single cloud for all their diverse IT needs, it would be among the largest pools of IT resources ever built. And if cloud computing is an opportunity created by economy of scale, then operators will be the leading contenders to provide it.

This story works just as well in reverse. Network operators worldwide have listed cloud computing as one of their top three applications to create new revenues. A decision to offer cloud computing services and create infrastructure for these services -- as Verizon did by buying cloud provider Terremark -- will create an IT platform that can then be exploited for the operators’own IT applications. You could consider this a “service-first” cloud evolution.

The fact that service provider clouds obtain optimum economies of scale by serving many different missions means that their design has to support a range of uses that most private or even public cloud infrastructures could elect to avoid.

An example is the classic “IaaS versus PaaS versus SaaS” debate. Operators that want to host SaaS providers’ services may need to offer their partners IaaS or PaaS clouds, or both. Most cloud providers elect to support one approach, but a network operator’s cloud will likely have to support all three, since they offer customers different functions:

• IaaS services are the baseline cloud offering due to Amazon’s EC2 popularity.
• SaaS services are mandatory for small business and consumer offerings
• PaaS services are essential for some enterprise data center workload overflow and backup applications.

In short, all models are needed.

Operators turn to PaaS as service delivery platform architecture

Many network operators are now conceptualizing their own internal use of cloud infrastructure in PaaS terms as well. If we were to describe the architecture of a service delivery platform in modern terms, we’d call it a PaaS host because it combines a hardware platform with a structured set of middleware that creates a uniform development environment. Feature-building missions for the cloud are certain to be supported by assembling and even creating custom middleware.

For operators that want to offer app stores, developers are likely to operate inside a micro-PaaS sandbox that includes tools to manage the offerings and integrate them with operator billing and support applications.

IT vendors customize cloud middleware for providers

Operators have many choices of cloud architecture, and some of the popular IT giants are now customizing cloud middleware into packages for use by providers as well as enterprises. IBM’s Cloud Service Provider Platform (CSP2) and Microsoft’s Azure Platform Appliance are examples of this trend, and the use of an integrated toolkit has the advantage of ensuring compatibility and management commonality across the elements of cloud infrastructure. Some operators are also considering clouds created from open source or commercial dual-licensed software like Eucalyptus (Eucalyptus Enterprise Edition) and Hadoop’s Cloudera

.

The key to successful cloud computing infrastructure

For operators, the key to creating successful cloud services and supporting internal IT on the same infrastructure is ensuring that an IaaS framework built on data center virtualization can be extended upward first to the platform and then to the services level without creating voids in the management processes, and without compromising a single-resource economy of scale value proposition.

In the final analysis, the combination of the economies of resources and management efficiency will define the provider cloud and differentiate it from other cloud computing offerings in the market.

Author: Tom Nolle

Building multiple cloud services from a common infrastructure

“e Pluribus Unum,” which translates to “from many, one,” is a familiar phrase written on U.S. coins. For service providers planning cloud computing services, the motto and challenge is exactly the opposite: How can they produce many services out of one infrastructure, or “e Unum Pluribus”?

Most service providers have neither the option nor the desire to pick out a single model of cloud service and focus investment on it alone. Leveraging their infrastructure and operations to produce superior economies of scale is critical, and from a sales perspective, it will be easier to exploit the cloud quickly when enabling multiple cloud services rather than having only one.

Technically, cloud computing is a resource-pool strategy. A set of servers and storage devices located in one or more data centers form a pool of resources that are allocated to service customers on an as-needed basis. The larger the resource pool, the greater the efficiency and lower the unit cost of computing and storage. A lower cost base then permits network operators to offer computing and Storage as a Service at prices compelling to the buyer and at margins profitable to the seller.

The value of generality at the cloud service level is clear: More services equal more sales, which equal more resource-pool efficiency and higher profits. But the question is how to achieve the multiservice goal with a practical set of cloud computing tools.

Analyzing the components that enable multiple cloud services

The strongest starting point to answer the question of how to achieve multiple cloud services using one set of cloud computing tools is to consider how the cloud computing service in question—Infrastructure as a Service (Iaas), Platform as a Service (PaaS) or Software as a Service (SaaS)—would appear to a buyer and what the cloud resource interface needs to resemble. The goal of a universal infrastructure would be to offer all of those service appearances and interfaces at comparable efficiency and with full management capability. Where the requirements compromise the ability to use one infrastructure, operators need to weigh the cost and benefits of that offering. Here’s a closer look at the requirements of the three cloud services.

Infrastructure as a Service. IaaS services would logically appear to the buyer as a virtual machine host that is essentially an extension of the buyer’s own data center virtual resource pool. Running an application on a virtual host means creating a machine image that includes the application and its operating system software and middleware, and then loading it on the selected cloud server.

Since the application image and operating software is provided by the user, the cloud provider has limited options for managing the internal behavior of the application. The virtual machine host can be managed but not the machine image software. Still, this application is what made Amazon’s EC2 famous, and operators almost universally expect to support it. A virtualized set of servers and a director function are required to assign free virtual machines to the best available server resource.

Platform as a Service. PaaS services can be viewed as an IaaS framework that hasdefined one or more middleware “services” that are available to the applications within the cloud. Therefore, these applications are not included in the machine image provided by the user. The most common platform service is Database as a Service, followed closely by management services.

To the user, the value in consuming a platform service versus having a completely customer-provided machine image of the application is that the platform service is cloud-aware and can be managed and optimized for a cloud-hosted execution. Cloud databases are almost mandatory for applications that will involve data exchange among application components. Otherwise, these components would have to develop an internal mechanism for data sharing that could be relied upon to work efficiently within a cloud, where resource-to-application assignment is nearly invisible to the user. Cloud management tools, including tools to manage application performance, allow both users and the cloud provider to merge application performance issues with virtual machine resource performance, giving a better management picture.

As applications begin to consume platform services, the IaaS framework is likely to take on the appearance of a data center optimized for service-oriented architecture (SOA). The idea of the user providing a “machine image” is replaced by the idea of the user providing an application to catalog—an application that will be represented in an SOA directory (UDDI) and instantiated on demand, using the provider’s virtual machines. This represents PaaS services as effectively an SOA overlay on IaaS infrastructure, a model that seems explicitly adopted by Microsoft with Azure and by IBM with its Cloud Service Provider Platform.
Platform services are represented as SOA services and accessed as any internal software service would be accessed. This also facilitates the integration of PaaS cloud services with enterprise software; any SOA tool can provide the binding required.
Software as a Service. Moving up the stack to SaaS, we again find that the next level of cloud services can begin as an extension of the last. The SOA abstraction of a “service”can be applied both to application components and to entire applications. In the former case, a SaaS offering would be associated with a SOA-compatible interface (including Simple Object Access Protocol, SOAP), and could be registered as a service in a SOA directory for access as described with a PaaS cloud service. In this case, an enterprise might be consuming a cloud SaaS component within an application otherwise hosted in-house.
If the entire application were to be represented as a service, then the SaaS offering could be made through a traditional URL or RESTful interface and be accessed by a browser. This would be the type of SaaS service that well-known Salesforce.com provides. Operators could provide these types of services directly (using open source software, licensed software or software they develop), or they could offer wholesale PaaS or IaaS support to third parties that would then offer SaaS services.\

Building more complex cloud services on the components of the others
Cloud services at the higher levels, which include SaaS and PaaS, displace more user support and license costs and, therefore, justify higher prices. If operators can use the same infrastructure to offer these services as they use for the more general IaaS services, they can increase their profits and provide cloud services more easily integrated with enterprise IT. That can be particularly valuable in addressing the data center backup and application overflow opportunities that enterprises value most highly for cloudsourcing.

Author:Tom Nolle

A cloud provider's guide to building multiservice cloud platforms

Following the initial excitement that cloud computing will change life as we know it, we’re now in the heavy-lifting phase that — over time — will make cloud services a trusted reality.

But to help enterprises, organizations and government agencies embrace the cloud for mission-critical applications rather than only for an economical test and development platform, cloud providers have to step up in a big way to design the right infrastructure, whether they’re entrenched telecom providers, newer cloud specialists or niche providers

Challenged on many fronts to usher in the era of scalable and reliable cloud services, providers have to design an architecture and build cloud platforms that can accommodate multiple services without losing the economies of scale that made cloud services an attractive proposition in the first place.

This expert lesson on building multiservice cloud platforms like a pyramid scheme (but in a good way), by frequent contributor Tom Nolle, looks at a number of issues all cloud providers have to address, including capitalizing on a common cloud platform to build multiple services, using your cloud platform for internal and purposes and getting your cloud database strategy in order to facilitate application performance.

Here are the building blocks you need to get your cloud platform strategy up to speed.

1. Building multiple services to operate on one cloud platform

First and foremost, most cloud providers don’t have the luxury of building a cloud platform for each service they plan to offer, which means that to make the economics work for themselves and for their customers, they need to leverage their infrastructure to build multiple services to operate off one cloud platform. Using your resource pool of data centers, servers and storage devices wisely is the key to building more complex cloud services. Providers can start with basic cloud platform elements needed for Infrastructure as a Service, then build more complex cloud services including Platform as a Service, Software as a Service and beyond, on top.

Brush up on building multiple services off one cloud platform.

 

 

2. Building a cloud computing infrastructure to serve dual purposes

Moving forward, cloud providers have to remember that they are giant IT consumers themselves, and they need to benefit from their cloud infrastructure as much as any customer. Thinking outside the last-gen silo mentality, cloud providers need to move beyond supporting multiple cloud services for customers and make sure their cloud platforms function for their own internal OSS/BSS and internal IT needs. Any provider’s cloud platform has to fill the dual roles of internal IT and customer services, which means using Infrastructure as a Service architecture built on data center virtualization to benefit from the economies of scale.

Understand the dual-purpose cloud platform for internal and external use.

 

 

3. Cloud Database as a Service: Planning your DBMS strategy

To offer cloud services, cloud providers have to have an effective database strategy—otherwise customers won’t be able to get to their data. So why not consider rolling out Database as a Service at the same time, since you’ll need a cloud database strategy that won’t affect application performance anyway? Database as a Service is a good differentiator for cloud providers, but it requires careful analysis of your cloud infrastructure, your storage service model and your database management system models.

Read this article to find out if Database as a Service should be in your future.

Author:Tom Nolle

Time for Cyber Discourse on China

China is being accused of hacking corporate, government and military networks in the U.S. for economic gain.  Policy makers need to be versed in cyber security and figure out how to respond.  

JAMIE METZL CAUSED quite a stir late this summer with an article he wrote for the Wall Street Journal in which he blasted China’s computer hacking efforts. Metzl, executive vice president of Asia Society and a former higher-up in the State Department and National Security Council, condemned China’s actions as “running roughshod over global norms” to advance its economic interests.

Unfortunately for him, he used McAfee’s Shady Rat research—which received criticism from several experts in the industry—as the backbone for his diatribe against China. Regardless, the bigger point here is China’s 10 percent annual economic growth, a staggering number according to bean counters, isn’t exactly being built solely on blood, sweat and tears Metzl and others we’ve talked to and listened to say China is relentless in its efforts to steal intellectual property, trade and corporate secrets, and anything else that will give them an economic edge—or growth spurt. I’ve had more than one casual conversation land on the topic that some product some startup has been slaving over suddenly shows up on the China market months ahead of a potential launch here.

 Are we covering new ground here? No. But it’s worth reminding those who will listen that the Chinese are on our networks and are leveraging state-sponsored or politically motivated computer hackers to steal anything that isn’t nailed down. China’s efforts aren’t limited to big business either. Despite Art Coviello’s best efforts to tap dance around the obvious, I’ll take some journalistic license to read between the lines and conclude the Chinese were behind the SecurID attack. The attacks that compromised the company’s flagship SecurID authentication technology have been the security story of the year. The seriousness of the attacks quickly came to light when it was revealed they were merely a jumping off point for a downstream attack on the defense industrial base as Lockheed Martin and others subsequently reported they too had been breached.

China computer hacking is also the suspected culprit behind the Aurora attacks on Google, Adobe and upwards of 20 other enterprises, manufacturers and defense contractors in 2009. Plus, two Department of Defense reports released in the last 20 months name China as active in moving digital assets off American networks—corporate, government and military. Can we stop the politically correct pretense and examine closely in public circles the impact of these intrusions upon our economy and national well-being? Granted, if we cast that spotlight on the Chinese, we’re likely to get an equally bright light shined upon U.S. activities in China, Iran (hello Stuxnet) and other foreign interests. So be it. It’s time for ground rules and time to tame the Wild West before real lives are lost, not just nuclear centrifuges and software source code. There needs to be discourse at a policy level in Washington on cybersecurity and a clear understanding from legislators on these activities and their ramifications. The call for “offensive” weapons in cyberspace is also rattling around offices at the NSA and DoD and clearly some have been developed (hello again Stuxnet), but there are no rules of engagement written in stone yet in terms of how to react and reply to cyberattacks. How long before a physical, military response from either side follows up a cyberattack perpetrated by either side without a means for attribution of the attack or channels of communication between policymakers well versed in cyber? The Chinese aren’t shy about taking land or IP by eminent domain it seems. Pretty much anything is in scope to advance their economic agenda, according to Metzl’s op-ed in the Journal. If so, it’s time to bring cyber to prominence in Washington and internationally begin some real forward thinking before real companies are unable to compete in their respective markets, or worse, real lives are lost. BY MICHAEL S. MIMOSO

Cloud computing server architecture: Designing for cloud efficiency

A server is one of those industry terms whose definition is broadly understood yet at the same time ambiguous. Yes, "server" means a computing platform on which software is hosted and from which client access is provided. However, the generalizations end there. Not only are there many different vendors that manufacture servers, but there are also a variety of server architectures, each with its own requirements. A mail server, a content server, a Web server and a transaction server might all need a different mixture of compute, network and storage resources. The question for many providers is: What does a cloud computing server need?

The answer will depend on the target market for the cloud service and how that market is reflected in the applications users will run. Servers provide four things: compute power from microprocessor chips, memory for application execution, I/O access for information storage and retrieval, and network access for connecting to other resources. Any given application will likely consume each of these resources to varying degrees, meaning applications can be classified by their resource needs. That classification can be combined with cloud business plans to yield a model for an optimum cloud computing server architecture.
For a starting point in cloud computing server architectures, it's useful to consider the Facebook Open Compute project's framework. Facebook's social networking service is a fairly typical large-scale Web/cloud application, and so its specific capabilities are a guide for similar applications. We'll also discuss how these capabilities would change for other cloud applications.

Cloud computing servers needs may not align with Facebook Open Compute
The Open Compute baseline is a two-socket design that allows up to 12 cores per socket in the Version 2.x designs. Memory capacity depends on the dual inline memory modules (DIMMs) used, but up to 256 GB is practical. The design uses a taller tower for blades to allow for better cooling with large lower-powered fans. Standard serial advanced technology attachment (SATA) interfaces are provided for storage and Gigabit Ethernet is used for the network interface. Facebook and the Open Compute project claim a 24% cost of ownership advantage over traditional blade servers. Backup power is provided by 48-volt battery systems, familiar for those who have been building to the telco Network Equipment Building System (NEBS) standard.

The Open Compute reference has a high CPU density, which is why a higher tower and good fans are important. However, many cloud applications will not benefit from this high of a CPU density for several reasons:

• Some cloud providers may not want to concentrate too many users, applications or virtual machines onto a single cloud computing server for reliability reasons.
• The applications running on a cloud computing server may be constrained by the available memory or by disk access, and the full potential of the CPUs might not be realized.
• The applications might be constrained by network performance and similarly be unable to fully utilize the CPUs/cores that could be installed.

If any of these constraints apply, then it may be unnecessary to consider the higher cooling potential of the Open Compute design, and shorter towers may be easier to install to support a higher overall density of cloud computing servers.

How storage I/O affects cloud computing server needs
The next consideration for cloud computing server architecture is storage. Web applications typically don’t require a lot of storage and don't typically make large numbers of storage I/O accesses per second. That's important because applications that are waiting on storage I/O are holding memory capacity while they wait.

Learn more about how virtualization technology offers cloud infrastructure flexibility.
Check out our primer on identifying the most flexible cloud technology models.
Find out how new services, such as cloud computing, are driving changes to telecom data center architecture.

Consider using larger memory configurations for cloud applications that are more likely to use storage I/O frequently to avoid having to page the application in and out of memory. Also, it may be difficult to justify the maximum number of CPUs/cores for applications that do frequent storage I/O, as CPU usage is normally minimal when an application is waiting for I/O to complete.
A specific storage issue cloud operators may have with the Open Compute is the storage interface. Web applications are not heavy users of disk I/O, and SATA is best suited for dedicated local server access rather than storage pool access.
Additionally, it is likely that a Fibre Channel interface would be preferable to SATA for applications that demand more data storage than typical Web servers-- including many of the Platform as a Service (PaaS) offerings that will be tightly coupled with enterprise IT in hybrid clouds. Software as a Service (SaaS) providers must examine the storage usage of their applications to determine whether more sophisticated storage interfaces are justified.
Cloud computing server guidelines to consider
Here are some summary observations for cloud providers looking for quick guidance on cloud computing server architecture:

• You will need more sophisticated storage interfaces and more installed memory, but likely fewer CPUs/cores for applications that do considerable storage I/O. This means that business intelligence (BI), report generation and other applications that routinely examine many data records based on a single user request will deviate from the Open Compute model. Cloud providers may also need more memory in these applications to limit application paging overhead.
• Cloud providers will need more CPUs/cores and memory for applications that use little storage-- particularly simple Web applications -- because only memory and CPU cores will limit the number of users that can be served in these applications.
• Pricing models that prevail for Infrastructure as a Service (IaaS) offerings tend to discourage applications with high levels of storage, so most IaaS services can likely be hosted on Open Compute model servers with high efficiency.
• PaaS services are the most difficult to map to optimum server configurations, due to potentially significant variations in how the servers will utilize memory, CPU and especially server resources.
• For SaaS clouds, the specific nature of the application will determine which server resources are most used and which can be constrained without affecting performance.

The gold standard for server design is benchmarking. A typical mix of cloud applications running on a maximum-sized, high-performance configuration can be analyzed for resource utilization. The goal is to avoid having one resource type -- CPU capacity, for example-- become exhausted when other resources are still plentiful. This wastes resources and power, lowering your overall return on investment (ROI). By testing applications where possible and carefully monitoring resource utilization to make adjustments, cloud providers can sustain the best ROI on cloud computing servers and the lowest power consumption. That's key in meeting competitive price points while maximizing profits.

About the author: Tom Nolle is president of CIMI Corporation, a strategic consulting firm specializing in telecommunications and data communications since 1982. He is the publisher ofNetwatcher, a journal addressing advanced telecommunications strategy issues.

Four ways to reduce data center power Consumption

At a recent Association of Information Technology Professionals data center panel discussion, a seasoned group of IT admins discussed meeting customer power demands, with the consensus that demand is insatiable. Even as budgets seesaw from abundant to sparse, the demand curve never flattens, instead climbing skyward. The Jevons Paradox, the nineteenth century axiom of “the more we produce, the more we consume,” looms large in IT for the foreseeable future. Or as my colleagues say, “If you build it, they will fill it.”  The first panel warning was that virtualization is not a cure-all for reducing data center power consumption. Of course, there’s a clear advantage to high-density computing --cramming many virtual machines (VMs) into a single server -- but CPU demands for power and cooling still grow with each VM. In many cases, power and cooling costs shift from distributing power across lots of small servers to boosting power to cool red-hot VM-hosting systems.

 It’s not just about CPUs cranking out BTUs, which raises the next issue. The power needed for cooling, lighting, battery backup (UPS systems) and other environmental factors usually accounts for 35% or more of a data center’s total energy consumption, regardless of how efficient a building is built. Servers gobble watts, and keeping them happy is major overhead.

 Another important panel consideration was to shorten the return on investment. Returns have to show up fast -- within days or weeks. Nevermind three-to-five year returns -- in this economy, strained budgets can’t wait. All of the panelists insisted that IT managers have to show fast results before selling long-term solutions.

 So what are fast turnaround projects that deliver results quickly? The suggestions below are somewhat small, mostly single project efforts, or at least quick changes without high infrastructure costs. Combining them could create a synergy where the sum is greater than the parts, but doing so isn’t required to make efficiency gains. There were four main ideas presented for reducing data center power consumption, each of which can be implemented separately.

Switch to variable-speed fans

Recent research found that power consumption drops 30% for every 10% reduction in fan speed. As the name implies, these fans only consume power when needed, only running at the speed required, based on fairly sophisticated thermostatic measures. Since these fans slow down over long periods of time at low CPU utilization, they quickly decrease powerusage with each non-turning blade. And don't stop with servers; check cooling features of UPS devices and power supplies of various appliances on the same power grid, plus any other hot spots that may have a fan spinning for a while.

Raise the air temperature      

According to data center infrastructure suppliers, modern servers can perform well up to 77 degrees Fahrenheit. Yet many data centers have cooled servers down to mid-60 degrees F for years. By raising the ambient air temperature a few degrees, there can be an immediate drop in power usage by the cooling system with no server performance impact. There’s no overhead or investment needed, although close monitoring and a solid pilot program would be advisable to avoid unpleasant surprises. Granted, a slightly warmer server room can be a disconcerting change. For example, the dress code may have to be adjusted to allow for lighter clothes in warmer conditions.

             Use bigger, slower drives

 Of course, this should not be done for high-demand transactional processes, such as financial databases or critical 24-hour systems. But by delegating a percentage of mostly unused files to a lower tier of storage, big, low-energy demand drives can replace small, fast units. In turn, less drives burn less energy, creating less heat. This can be an expensive undertaking, but as most shops build out more storage every quarter, they should see it as a worthwhile investment.

Use hosted services                

Although moving IT workloads to a cloud or colocation provider externalizes the carbon consumption off to the host site, many will concede that big vendors are experts at squeezing the most out of a kilowatt. By using hosted services, you’ll be able to focus on delivering better value at a lower cost for your customers.

The risks of data center power consumption projects

IT organizations need to acknowledge the inherent risks in energy-efficiency projects. As one power company director put it, in a high-density, highly efficient environment, the data center can go thermal in seconds. Several recent high-profile outages started as a partial interruption, but cascaded to bringing down the entire facility. The catalyst – overheating that spread from rack to rack until all systems shut down for self protection.

The final warning: Spell out any risks before implementing changes to the data center and make sure to get executive support before pursuing any of these tactics of reducing data center power consumption.

By Mark Holt, Contributor, SearchDataCenter.com

Aberdeen Helps Out The U.S. Army With Its Stirling Storage Solutions

Storage is an important facet of any data center, but when the data center in question belongs to the U.S. Army, finding a secure, reliable storage solution is even more vital. With the U.S. Army in need of a large, fast storage solution for hosting user data and providing disk-based backups, Jeff Dupere, network administrator of the AATD (Aviation Applied Technology Directorate), turned to Aberdeen. “We also needed a complete turnkey solution to begin leveraging a virtualized environment,” Dupere says. “This included a storage-area network, hosts for running the virtual machines, and the Fibre Channel connectivity to tie the two together. Also, the entire solution needed to be VMware-certified.”

The Right Stuff

Aberdeen’s sales rep provided Dupere with the expertise of one of Aberdeen’s engineers, who helped to determine the best solution for the U.S. Army’s environment and needs. The solution included four Stirling 266 servers (2U SuperServers that support Intel’s Nehalem processors), an XDAS D-Series SAN (a 4U, 24-bay Fibre 8G/SAS 6G DAS), and a Stirling X888 (an 8U storage server). “The Stirling 266s provided us with the hosts for running our VMs, the DAS gave us the backend storage for hosting the data for those VMs, and the X888 met the storage requirements for our user data and disk backups,” Dupere says. The equipment offered all of the required features necessary for the job, such as full redundancy, high amounts of storage, and fast I/O.

Aberdeen’s solution was evaluated against several well-known competitors. “In the end, the solution Aberdeen offered was well-supported, exceeded our performance requirements, and was significantly less expensive than the other offerings,” Dupere says. “They are also one of the few companies I’ve seen to offer a full five-year warranty on virtually all of their equipment at no additional cost. Given our life cycle commonly exceeds three years on server hardware, this was a significant advantage.”

Stirling 266

With its dual LGA 1366 sockets, the Stirling 266 is a 2U SuperServer 6026T-TF that’s capable of utilizing Intel’s high-end Nehalem processors. There are also 12 DDR3 sockets that can support up to 192GB of Registered ECC DDR3 or up to 48GB of unbuffered memory with ECC. Helpful features include support for KVM over LAN and virtual media over LAN through the integrated IPMI 2.0 and Realtek dedicated LAN. The U.S. Army will also enjoy peace of mind with built-in PC health monitoring. There are four onboard voltage monitors for the processors, as well as six fans with tachometer status monitoring. Environmental temperature monitors, including chassis and CPU  overheat alarms, provide further security. “[The Stirling 266] is what the engineer recommended, and based on previous experiences, I was happy to go with his suggestion,” Dupere says. 

Aberdeen also preloaded ESXi on the hosts of the server, which made the environment turnkey for the Army. “We more or less just had to install the equipment in the racks and turn it on,” Dupere says. “They also were able to provide all the extras we needed (additional drives, controllers, etc.) to make onsite repairs much quicker.”

Stirling X888

The Stirling X888 storage server can provide up to 100TB of storage, and the U.S. Army took advantage of the X888’s dual SFF-8087 miniSAS connectors to connect with the Aberdeen XDAS D-Series SAN to deliver up to 196TB of storage. “We needed a ton of storage and very fast I/O,” Dupere says. “The Stirling X888 had this in spades—so much so that we’re in the process of buying another to supplement our environment.” For fast I/O when needed, the server features quad Gigabit Ethernet LAN and dual SAS expansion ports. And Aberdeen’s Teaming Technology offers transfer rates up to 430MBps with an added XDAS-iSCSI RAID enclosure. The Army also benefits from Intel’s QuickPath Interconnect Tech-nology, which can be found on Intel Xeon 5500 Series processors, providing them with 6.4GTps and 4.8GTps data transfer speeds.

The SAS RAID on the X888 includes dual IOP348 1,200MHz PCI-E controllers, and each controller features 512MB of DDR2-553 SD RAM with ECC protection. The controllers support RAID 0, 1, 5, 6, and 10. Overall, the Stirling X888 can provide up to 1,200MBps of internal transfer speeds. The RAID controllers can also support SATA disk drives and SAS hard drives at the same time. External SAS connectivity is also available via the SFF-8088 connector. The miniSAS backplane on the X888 provides the Army with 50 (48 front, 2 rear) hot-swap/hot spare SATA 3.0 drive bays.

The 8U chassis comes with a 1,760W 3+1 redundant hot-swap power supply. Fan maintenance is also a breeze with the eight hot-swappable 80mm cooling fans. Options for flexibility with the build include a DVD-R or CD-RW and software upgrades such as iSCSI, NAS, SAN, and backup software.

Aderdeen XDAS D-Series

The XDAS D-Series is a 4U high-speed, high-availability SAN that features four 8GB Fibre Channel host ports on each of the controllers, which are ideal for the fast throughput and I/Ops needed by the U.S. Army. The XDAS also features full support for 6Gbps SAS drives to provide support for today’s fastest hard drives. The XDAS D-Series is also designed to be always available with fault-tolerant hardware modules, including redundant controllers, PSUs, and fans. As such, there’s no single point of failure for the U.S. Army to worry about.

Other helpful protection features include real-time problem detection and notifications through multiple monitoring capabilities on the XDAS. And intelligent firmware helps to protect against hardware failure to optimize performance and maintain data integrity. The XDAS D-Series uses a power supply that’s more than 80% efficient, and it can spin down the drive to save energy when the disks aren’t in use.

The U.S. Army benefits from the XDAS D-Series’ local replication abilities. XDAS storage provides the Army with both snapshot and volume copy/mirror capabilities. Full data copies let administrators quickly restore service if a RAID volume fails, and files can be restored or rolled back through the XDAS’ snapshot copies. Some upgrades the U.S. Army chose to add include increasing the bandwidth across the Fibre Channel fabric and adding larger drives.

Dupere says that the set of products the Army chose from Aberdeen has met or exceeded all of the group’s expectations. “Aberdeen is not a huge company like some of their competitors, as evidenced that few folks will recognize the name when you bring it up,” Dupere says. “That said, they offer solutions and equipment that are every bit on par with the commonly referenced brands in this genre. From sales, to the product itself, to the support you receive afterward, you will not be disappointed.”

A number of intelligent storage systems, including the Stirling 266 servers, Stirling X888 storage server, and the XDAS D-Series SAN, designed to offer high-capacity, high-performance solutions to organizations in need of fast and reliable storage.

“[Aberdeen offers] solutions and equipment that are every bit on par with the commonly referenced brands in this genre. From sales, to the product itself, to the support you receive afterward, you will not be disappointed,” says Jeff Dupere, network administrator of the U.S. Army’s AATD (Aviation Applied Technology Directorate).