Digital Walkabout

Private cloud to public cloud migration gets easier, but obstacles remain

Migrating applications from a private data center to a public cloud can be a rocky road, but many cloud customers choose to make their move without the assistance of cloud migration tools.

Cloud migration tools can help users move workloads from private to public clouds, as well as between disparate public clouds, but there are limitations. In some cases, these tools are required for management at the beginning of a cloud migration process; and in others, users doubt a mass-marketed tool can address all of the idiosyncrasies of their particular environments, such as custom applications.

"The tools will only address a very small number of the needs that we have, and for the amount of work that goes into learning how to use them, being an engineering-centric organization, it's easier for us to build our own tools," said Brian Hansen, senior director of operations for Tango.me, a free mobile voice and video call company. It bases some of its infrastructure on Amazon Web Services (AWS) and some on a private data center virtualized with VMware Inc.'s vSphere. A third-party tool "would have to be amazingly better" than his in-house tools to win his business, Hansen said.

Though Amazon offers an import tool for VMs, manual redeployments of applications are often performed during migrations from private data centers to AWS. Software provider Bluebird Auto Rental Systems, for example, recently moved 200 databases for its car rental customers between Equinix-hosted private data centers to the AWS public cloud, according to Vice President Phil Jones.

Thrifty UK was the biggest database moved, with some 320 users total. It was a process that went on for about a month and consisted of manual checks to make sure the environment was right and all the software loaded, Jones said. "We did a couple of dry runs with a backup of their database; they did load testing to make sure there wasn't going to be any performance problem on Amazon, and then we set a date to switch them over, so we moved the database a second time," he said.

While this was a straightforward process for the IT pros at Bluebird, it isn't for everyone. More than half of the respondents to a recent TechTarget CloudPulse survey said that manual migration processes are a major hurdle to adopting cloud.

Of 1,297 respondents, 52% said that the biggest challenge to public cloud adoption is application suitability, which forces businesses to rewrite or convert workloads. A lack of interoperability or integration with the cloud provider was also cited as a challenge by 32% of respondents.

Indeed, there are real-world challenges to migrations between private data centers and public clouds, according to Scott Sanchez, director of strategy for Rackspace, which is working on software to ensure workload portability between private and public OpenStack-based clouds this year.

"Enterprises need to have some flexibility around how their networks are configured; they need to have some flexibility about what hardware they purchase, because they all have preferences," Sanchez said.

An enterprise data center may use different storage than a cloud provider, for example. The IP and network infrastructure might be configured differently. These differences present migration challenges unless specific software modules are written into OpenStack to abstract the storage and networking layers of the infrastructure, regardless of type or vendor, Sanchez said.

"We've got portability now between our Cloud Zones and public cloud because we're able to configure them the same, but as we get into a private cloud, part of the balance we have to strike is, where do we draw the line in terms of what customers can change and what they can't change?" he said.

Cloud migration tools lift some burdens, but not all
VMware-based tools like vCloud Director and vCloud Connector can be used to move workloads between VMware-based private and public clouds, but they aren't currently compatible with other hypervisors.

Bryan Bonds, senior systems administrator for eMeter, an energy information division of Siemens located in Foster City, Calif., said his company has recently engaged with vCloud service provider Bluelock, and uses vCloud Director and vCloud Connector to move workloads from a private data center and Bluelock's public cloud service.

From vCloud Director, the system to be moved is shut down, and from within the operating control panel, an administrator will tell it to move the system, and select one of the company's virtual data centers from a list as a destination. With vCloud Connector in place, one of the things that appears on that list is the Bluelock environment. "It probably takes us half an hour of actual work to move a system from one side of the country to the other, and we turn it on and then we're done," Bonds said.

That is, if the system isn't too large to conveniently be moved across a WAN link. Systems of 100 GB or more are usually copied to an external hard disk drive and shipped to Bluelock, where staffers then import the virtual machine manually, Bonds said.

Tools like CloudSwitch also promise automated migration from private to public clouds, as well as between disparate public clouds. Verizon-Terremark, owners of CloudSwitch, said that the tool has helped customers in the financial services and health care industries move workloads from cloud to cloud, whether private or public.

RightScale and Racemi are two more vendors that offer cloud migration tools. RightScale can be used to deploy virtual machines quickly in the cloud, and can also apply a multi-cloud image beneath virtual machines to be moved from cloud to cloud. Racemi dubs itself "the moving company of the cloud."

Racemi, which is primarily marketed to cloud service providers, starts at a subscription price of $299 for the first four migrations, and charges $249 for the next five; VMware's vCloud Suite, including vSphere Enterprise Plus, vCloud Director and vCloud Connector, starts at $4995 per CPU license; and CloudSwitch starts at $25,000 for an annual license.

 

Mobile Cloud - Apps let enterprises handle the risks of Cloud

The overwhelming need to provide centralized management and security for multiple types of mobile devices is driving the increased adoption of mobile cloud computing and mobile cloud applications.

This is especially true for organizations adopting a bring your own device (BYOD) strategy. Over time, mobile devices will be able to access many of the traditional mission-critical business applications.

The growing acceptance of the mobile cloud has led to several notable trends in IT. Here's a rundown of those trends, the advantages and disadvantages of cloud computing and some thoughts on how they might affect organizations going forward:

Cloud trends in mobile applications
Gartner Inc. says that the mobile applications market is highly competitive and attracting the interest of many stakeholders. Gartner predicts that end users will spend about $15.9 billion on mobile applications in 2012, and those investments will in turn drive hardware sales and new technology innovation.

As the workplace becomes more mobile, with about 140 million smartphones and tablets entering companies in 2013 and more than 60% of enterprises endorsing BYOD strategies, custom mobile apps will emerge as a way for corporations to streamline how they do business.

A survey by AppCentral shows that about 19% of enterprises have in-production app stores. Enterprise applications stores are the direct result of BYOD strategies. Companies such as CDW and General Electric have implemented private stores to help take the burden off IT to move apps to individual devices or upload each application to a public app store.

Because more than 60% of corporations are implementing BYOD strategies, according to some estimates, many different types of mobile devices may be in use within a single corporation. It is too costly to provide native apps for all of these different types of devices, paving the way for mobile cloud apps that can be built once and used with many different mobile devices.

New technologies such as HTML5 and Sencha Touch are enabling mobile cloud applications developers to build apps with more polish. Mobile cloud apps have the big advantage that they can run on many mobile device platforms that have standards-compliant Web browsers and are much less expensive to build.

Mobile hybrid apps are being used in those cases where on-device features are needed but the mobile cloud app is the preferred mobile app choice. A hybrid app is one that is written with the same technology used for Web apps and mobile Web apps, but is hosted or runs inside a native container on a mobile device. A Web-to-native abstraction layer enables access to on-device features that are not accessible in mobile Web apps.

Independent retailers will begin creating public app stores. BestVendor and Listio are examples of two directory sites that could grow to compete with Apple.

Mobile cloud trends
In the near future, as developers move away from native apps, the world will see more consumer applications built as mobile cloud applications -- software that resides in the cloud but is accessed via an interface or browser on a mobile device.

The rise of mobile cloud applications can be attributed to the rising cost of deploying and supporting native applications and the fact that development tools are getting better all the time. HTML5 and CSS3 adoption is on the rise, making it easier to design and build mobile cloud apps. Basically, an HTML5 specification tells a Web browser what to display and a CSS3 specification tells the Web browser how to display it.

According to Gartner, corporate employees using smartphones and tablets for business purposes represent about 75% of the mobile cloud app market. The mobile cloud app market is expected to exceed $9 billion by 2014.

Corporations are getting pressure from their customers to build mobile cloud apps or make existing applications available via mobile devices. If they fail to do so, they are in danger of losing ground to competitors.

Mobile network operators (MNOs), such as Vodafone and Verizon, and other cloud vendors such as Global One are entering the mobile marketplace with mobile clouds that compete with other non-MNO vendors.

The resource-poor nature of mobile devices is driving the adoption of mobile cloud computing. Mobile apps are becoming more powerful as processing power is offloaded from mobile devices to the mobile cloud. Mobile devices that are 4G LTE (Long Term Evolution) compatible, which support bandwidth intensive apps and reduce latency issues, are helping to open markets for mobile cloud apps.

Security is a continuing issue for adopters of mobile technology. Some companies believe the risk of using mobile devices to access and update corporate data is too high. As a result, some companies allow mobile devices read-only access to corporate data, but not write access. Other companies, such as some financial institutions, allow customers to update their accounts using mobile devices. Authentication mechanisms are available to control how mobile cloud apps use data, locking down sensitive proprietary data while making less risky data more available.

A new approach to security for protecting mobile devices from malware and intruders is being examined. With this approach, multiple mobile cloud protection engines (virtual servers) run security programs that analyze a device activity file that each mobile device periodically uploads to them. Each mobile device has a software agent program installed to create the activity file. This approach promises to provide much better security for mobile devices than with a mobile device running a single security program.

Mobile clouds centralize the management and security of mobile devices, placing mobile device users into a safer, more controlled environment. This is being driven by the plethora of mobile devices of differing types, with different operating systems.

Powering BYOD with Mobile Cloud

The bring your own device (BYOD) trend is growing, with more businesses allowing employees to use their own smartphones and laptops for work purposes, which presents a big opportunity for cloud service providers to sell the cloud as an alternative to storing company data on personal devices.

An increasing number of businesses are removing company-issued smartphones from their budget lists and allowing employees to use their own smartphones. As a result, IT departments are faced with the task of not only supporting employees' smartphones, but also making sure work-related files accessed on personal mobile devices remain secure.

Employees are pleased with the BYOD trend because it means they don't need to carry two laptops, two smartphones and two tablets. But cloud providers say IT departments often feel somewhat squeamish about allowing employees to use their own devices for work, because they need to ensure that company email, corporate secrets and sensitive data won't be shared with competitors if devices get lost, stolen or hacked, or if their owners move on to new jobs with competitors. And that's where providers say the cloud is increasingly starting to factor in as a solution.

"Businesses are looking at the cloud as a solution to BYOD because they'll have more control from both a security and disaster recovery standpoint," noted Robert J. Chandler, president and CEO of Cloud9 Real Time, a cloud provider based in San Diego. "Almost on a daily basis, a business will contact us and let us know that an employee has left the company and we need to wipe their information remotely and terminate access."

Getting comfortable with cloud apps on personal devices
Globally, almost 60% of full-time employees participate in some form of the BYOD, according to a research white paper released in November by Ovum and commissioned by cloud provider Logicalis. And almost three-quarters of the 1,242 IT professionals surveyed in 2012 by reseller and service provider CDW said their organizations' decision to take on cloud computing was "significantly influenced" by employees' use of personal mobile devices. Thirty-nine percent had already adopted a cloud solution, up from 28% in CDW's 2011 survey.

"The BYOD trend is not one that's going to be stopped," said Jeetu Patel, general manager of EMC's Syncplicity business unit. "It's not just BYOD; it's the multiple devices people own for different purposes. When I'm on a train, I want my iPad, at work I want a laptop and at home I want a big-screen desktop. And I want all of my data to follow me."

Many customers end up feeling more comfortable with containing corporate information on the cloud, rather than allowing employees to keep data on their personal devices.

For example, Jennifer Katrulya, CEO of Danbury, Conn.-based accounting firm Business Management Resource Group (BMRG), a Cloud9 Real Time customer, said all her employees work on their own smartphones and tablets. Employees share documents through a cloud application and are advised not to download and save sensitive work documents on their personal devices.

"We have a list of applications that we ask them not to keep on a device because we feel that the best way to maintain security is to keep things off the devices to begin with," she said. "We are careful to use applications in the cloud, and we've been pleased with the way it's been going."

Cloud security still a stumbling block in BYOD trend
Cloud providers say they are able to market mobile cloud services to businesses as the way to go, largely because they are able to show clients new features that allow managers to have greater control over the privacy of company information on employees' personal devices.

"We can have more control over things with the cloud," Patel said. "If I leave my iPad at the gym, I want to be able to make sure that I can remote wipe any sensitive data from it. If I send an email attachment, I want it to have an expiration date. And if I have a contractor that I am sharing data with, when the contract is up I want to be able to stop sharing that data with them."

Providers say IT departments often come in with long lists of requested features in an effort to ensure that data is kept secure when accessed from personal mobile devices. Patel noted that every organization's security standards are different and, as a cloud provider, he makes sure he adapts his company's cloud proposals to address each client's needs.

Yet providers do continue to face the challenge of selling the cloud to businesses that question whether cloud security is adequate. It can be tough to market the cloud as a BYOD enabler when clients are not sure the cloud is secure enough to begin with, said Dean Weber, chief technology officer for the cyber security group at CSC, a cloud and managed services provider based in Falls Church, Va. Businesses are somewhat reluctant to embrace cloud as a solution to BYOD, he added.

"Organizations are coming to us and wanting to move to the cloud and also saying they have a BYOD problem, but what they're not doing is putting the two of those together yet," Weber said. "I don't believe they are doing them simultaneously, for the most part, because organizations don't have that risk appetite."

Weber said that delivering cloud services in the precise, secure way companies require can be costly, which leads some businesses to shy away from adopting the cloud for BYOD purposes. Some companies are starting out by testing relatively low-risk assets, such as email, while keeping more precious applications out of the cloud.

Cloud service providers say they attempt to alleviate those concerns by recommending that their customers create BYOD policies that spell out exactly how personal devices can be used for work purposes, including what kinds of documents should be saved on the devices and which ones should be accessible via the cloud. In the research by Ovum and Logicalis, only 20% of those who were using their own devices for work purposes had signed a BYOD policy.

Patel noted that some businesses talk about locking down a majority of data from staff, providing a limited few the access to cloud applications and leaving the majority out. But as a cloud provider, he tries to tell potential clients that this way of thinking is a mistake.

"If you say you're not going to allow people to use Dropbox, that's the worst thing you can do," he said. "What people do is go create a separate network so they can use Dropbox. It's like telling your kid when they're a teenager, 'You should not do these 10 things' without explaining the reasons; they'll go out and do it just to rebel. Instead, the organization needs to be thinking that it can enable a better service for users, so people can get the service they want, while IT can get the security it wants."

 

OpenStack Grizzly release pushes scalability, high availability

New features in OpenStack deliver sought-after capabilities, but IT pros disagree on whether the updates are enough to make the open source cloud OS ready for prime time.

The latest release of OpenStack, dubbed Grizzly, includes several incremental changes that add up to better scalability and high availability, which some attendees wanted at last year's OpenStack Summit.

• On the scalability front, OpenStack Grizzly introduces a concept called Cells, which offers a way to manage a number of distributed OpenStack compute clouds under one interface.
• Decentralization of services within OpenStack clouds is also a theme with this release. A new feature called NoDB is an update to the architecture that decreases OpenStack's reliance on a central MySQL database for orchestration.
• Version 3 of the Keystone identity management application programming interface (API) changes the way identity management is done so it's a more distributed series of "handshakes" rather than having to go through one centralized identity server every time. This makes interactions between services faster, and makes OpenStack more scalable, as well.
• In terms of high availability, the Cinder block storage API has matured with this release, and there are now plug-ins from 10 vendors, including EMC Corp. and NetApp Inc., that allow for the creation of pools of clustered, highly available block storage.
• The Quantum API now offers five new plug-ins from vendors, including VMware Inc. (Nicira), Brocade, Cisco Systems Inc. and Big Switch Networks, for network virtualization.
  "Quantum in Grizzly has come of age," said Josh McKenty, an OpenStack foundation board member and CTO and co-founder of Piston Cloud, an OpenStack independent software vendor. "But we need to get to the point where all plug-ins are equal citizens, and get rid of any kind of reliance on [the open virtual switch]."
• Monitoring, which was another wish list item at the OpenStack summit last year, is going to have to wait till the next release and the debut of a project called Ceilometer.
• Microsoft Corp.'s Hyper-V is supported in this release. Previously, Hyper-V had been booted out of the OpenStack Foundation for a time -- but Foundation officials didn't speak in great detail about new VMware integration, which was talked up at last year's show.
• Documentation, another enterprise requirement for production deployment, has improved with the publication of the OpenStack Operations Guide.

OpenStack's incremental progress
The OpenStack Foundation says there are at least a hundred production OpenStack clouds at organizations that include Best Buy and Comcast.

Meanwhile, IT pros are split on whether the Grizzly release is really a tipping point for OpenStack. The focus on operational issues rather than brand-new features has John Treadway, vice president for Cloud Technology Partners, an IT consulting firm in Boston, convinced that OpenStack is maturing. "Now you don't have to deal with a single point of failure in your control layer," he said. "Things like that make it safer and more robust for enterprises and service providers to deploy OpenStack."

Meanwhile, another cloud consultant, Jared Reimer, co-founder of Cascadeo Corp., an IT consulting firm located in Mercer Island, Wash., said he's heard a lot of conversation among his clients around OpenStack, but hasn't seen a lot of action in terms of actual deployments. "Do I think OpenStack is going to be big news in the future? Yes," Reimer said. "Do I think Grizzly is big news? Honestly, no."

Don't let the bugs bite
As Grizzly comes to market, some experts are wary that the new version of the Keystone identity management API popped up relatively late in the development cycle, and point out that last time such a thing happened, the Diablo release in 2011 didn't work right out of the gate (because of a separate problem). Some changes to Quantum are also working their way through the bug fix process. "Most of the bugs get found when the broader community is working with the product, and we're still not at a point where it's that easy to deploy where we see a whole ecosystem deploying the early betas," said Piston Cloud's McKenty.

Every change to OpenStack goes through automated testing and then human review before any changes are merged into the codebase, said Jonathan Bryce, executive director of the OpenStack Foundation. "That's actually something that's gotten more rigorous with every release," he said.

OpenStack being an open source project means more transparency about bugs and glitches with new releases, Treadway pointed out. "Proprietary vendors also rush things to market," he said. "Open source means you know where the risks are."

Enterprises usher in mobile cloud computing

Mobile devices and mobile applications are changing the way IT organizations buy and use technology. Enterprise employees, not IT, are driving this trend. Even though cloud computing may facilitate end-user access to mobile apps, IT managers and developers must address a few nagging issues with mobile cloud computing.

Driven by the adoption of HTML5, increased mobile broadband coverage and the need for always-on collaborative services for the enterprise, the market for cloud-based mobile applications is expected to grow 88% from 2009 to 2014, according to Juniper Research. More than 240 million business customers will access cloud computing services via mobile devices by 2015; and that number could jump to 998 million mobile cloud users in 2014, reported ABI Research.

Mobile cloud computing is not mobile computing
In the past two years, mobile cloud computing has gained interest among a limited number of enterprises. Even though they may sound similar, mobile computing and mobile cloud computing are very different.

In traditional mobile computing environments, mobile applications run on a mobile device with the application data stored on the device. Running a mobile application has some advantages, most importantly, no latency or network bandwidth problems. But applications that run on mobile devices are generally limited in functionality and are not business-class applications.

Mobile cloud computing applications, however, run on servers that reside in the cloud. Application data also resides in the cloud and results are fed back to the mobile device over a network. This allows users to run more robust applications. There are problems, however, such as latency and network bandwidth issues for the transfer of data between the mobile cloud and the mobile device.

Two sides of the mobile cloud
There are at least two views of mobile cloud computing. One view is that of an enterprise IT organization adapting an on-premise, private cloud to connect various mobile devices such as smartphones and iPads or tablets -- where processing and data storage is on the cloud, not the device. In this instance, an enterprise user on a mobile device would access applications running in the private cloud through the device's browser.

A second view of a mobile cloud is referred to as mass-market mobile cloud. Mass-market mobile clouds involve players we do not generally associate with traditional enterprise clouds. These players include mobile network operators (MNOs) like AT&T, China Mobile, Vodafone, T-Mobile and Verizon Wireless, for example, as well as cross-network service providers (CNSPs). Another important component of mass-market mobile clouds is the application provider -- individual, application service providers and Software as a Service (SaaS) providers.

Previously, smartphone users were restricted to the mobile applications that their carriers provided in association with a cell phone company, such as AT&T Mobility and Apple. Mobile applications depended on the mobile device OS and other phone features.

The mass-market mobile cloud, however, allows users to access a wide variety of applications via their mobile device browser. A CNSP hooks up an MNO (to which a user is a subscriber) with an application provided by one of many individual application developers, ASPs and SaaS providers. The CNSP directs the MNO to charge the mobile device account for the application.

Enterprise clouds and mass-market mobile clouds appear to be quite different, but they are both based on common principles. Mass-market mobile clouds provide a way for enterprise staff to access mobile applications no matter what MNO or carrier they subscribe to.

Using mobile devices, enterprise staff can access applications in the enterprise cloud, provided the enterprise cloud is implemented to handle mobile devices and applications, or access them via their mobile devices from a platform owned and operated by a mobile cloud provider or CNSP. The CNSP enters into agreements with the individual MNOs and is responsible for providing access and payment services across multiple mobile network operators for applications.

What's good and bad about mobile cloud
Mobile cloud computing has its highs and its lows. Here are a number of advantages associated with mobile cloud computing:

• Users can share resources and applications without a high level of capital expenditure for hardware and software resources.
• Highly expensive and technical hardware is not necessary to run complex applications that run in the mobile cloud, which can reduce costs.
• Mobile cloud computing allows developers to have access to a larger market.
• Developers do not need to make a new build for each mobile device because processing and data is stored in the cloud.
• Applications run through a browser, so the end user's mobile device OS does not affect mobile applications. This allows developers to bypass restrictions that might be created using mobile operating systems.

But no technology is perfect. And while mobile cloud computing has its advantages, it's not without challenges. Many IT organizations must address the following issues directly or work with others in the enterprise to address them:

• Security: Internal communications no longer run over a private network; some run over less-secure public carrier networks.
• Mobile software: The software must run on multiple, frequently updated OSes and work with other mobile applications.
• Network availability and intermittency: A cloud application needs a constant connection; as mobile Internet capabilities improve, this will become less of an issue. HTML 5, which enables data caching through a mobile device, allows a mobile cloud application to continue working even if a connection is lost momentarily.
• Device division: Safely divides a single mobile device between professional and personal uses.
• Application development: Enterprises must determine whether to develop mobile applications internally or wait for key application vendors to come out with mobile device versions, such as table versions, of applications.
• Application downloading: Enterprises must determine if they should create an in-house enterprise mobile application store to ensure users download only enterprise-approved mobile applications.
• Remote management of mobile devices: Enterprise mobile device management tools enable IT teams to manage mobile devices by setting policies, configure devices to perform remote wipes, create enterprise sandboxes or secure virtual areas and keep personal data separate from corporate data using tools like encryption and passwords.

In terms of mobile device policies and processes, IT organizations have some experience with the widespread use of notebooks, for example. Compounding the difficulty is the large number of mobile devices (not just notebooks) that will be in use in the enterprise in the future. Some policies and processes will have to change and new ones will need to be added to meet these increased numbers:

What changes will need to be made to configuration management and management processes? What effect does mobile cloud computing have on the help desk? Old manual processes cannot keep up with the number of mobile devices employees use.

• Which mobile applications can be used for business?
• How will IT teams need to handle company-owned data on employee's mobile device if it's lost or stolen?
• How is company owned data on mobile devices partitioned from personal data?
• Who is allowed to use mobile devices?
• Who is responsible for technical support of mobile devices?
• Which policy is best for employee-owned mobile devices:
• End users sign consent forms in which they agree to have software installed on their devices to synchronize email, calendars, etc.
• Users grant permission to IT teams to allow remote wiping of lost or stolen mobile devices
• Users confirm they will adhere to enterprise-defined usage policies
• Will mobile device users be allowed to upgrade software without enlisting IT?

Don't hold back on mobile cloud
Today, with several mobile operating systems in the market, including Android and iOS, mobile application developers must pick and choose which OS to support, limiting their ability to broadly distribute applications. But mobile cloud computing is changing the way mobile applications are developed, acquired and used within the enterprise, which makes it easier for application developers to distribute apps to more users. Mobile cloud computing enables enterprise employees to use mobile devices to access data center resources that previously were only accessible via desktops, laptops and notebooks.

As mobile cloud computing use increases, enterprise IT managers wrestle with a few unanswered questions: What about security? Should employee-owned mobile devices be allowed to access company data? Should developers build mobile apps internally? What mobile device processes and policies need to be in place?

Despite these concerns, enterprises shouldn't "hold back" the use of mobile devices. Instead, they should determine how to take advantage of mobile cloud computing and plan for its explosion in order to remain competitive within their markets.

 

Cloud propels HPC sector

The HPC sector is showing enterprise IT what the public cloud can do: save gobs and gobs of money, while pushing the envelope on the scale of computation. But how likely is widespread adoption?

For a glimpse of the future of computing, the high-performance computing sector has always been a good place to look. In this age of cloud computing, HPC shops have been true to form, seizing upon cheap and plentiful public cloud resources to build ginormous compute clusters for short dollars -- and big results. In recent years, the worldwide HPC sector has been one of the lone bright spots in IT spending, estimated at $20.3 billion in 2011, and growing at a compound annual growth rate of 7.6%, according to IDC.

The availability of cheap cloud resources could propel the HPC sector still further, providing compute resources to a whole new class of users who were previously priced out, said Jason Stowe, CEO of Cycle Computing, a provider of HPC provisioning software. "Public cloud is democratizing access to utility supercomputing," Stowe said. "Before it was a have [or] have-not kind of environment." The trails blazed by HPC and public cloud could show enterprises with more modest needs how to take advantage of this new class of computing.

HPC in the cloud: The core of the matter
As with the TOP500, the yearly ranking of the world's fastest supercomputers, recent HPC cloud successes are nothing if not impressive. But whereas the TOP500 measures computing success in floating-point operations per second (FLOPS), cloud HPC projects tend to focus on cores.

Perhaps the most impressive project comes from computational chemistry firm Schrödinger, which created a whopping 50,000-core Amazon Elastic Compute Cloud (EC2) cluster to analyze 21 million drug compounds in just under three hours. The job -- a virtual screen that binds the compounds to a protein -- returned a list of more than 100 compounds that merited further testing.

That 50,000-core cluster featured many multiples more compute capacity than Schrödinger's internal resources, a 3,000-core cluster comprised of commodity 32-core nodes. "What we have available for a virtual screen is usually much less," which limits the scope of the research, said Rami Farid, Schrödinger's president. With smaller core counts, "you have to cut corners or it will not get done," he said.

In addition to the sheer size of the run, the Schrödinger job was notable for its low cost. Schrödinger paid AWS $4,900 per hour for three hours to run the job. Cycle Computing, which provided the cluster management software that enabled Schrödinger to export its jobs to the cloud, estimates the value of the EC2 infrastructure that ran the job at $20 million.

Other Amazon customers take advantage of spot instances to drive down the cost of running HPC jobs even further. Spot instances allow users to name the price they are willing to pay for unused EC2 capacity, and will run the job whenever the bid exceeds the current price, which varies based on real-time supply and demand.

A recent example is the University of Wisconsin's Morgridge Institute for Research, which completed more than 1 million core hours of gene indexing in just more than a week. The memory-intensive job required the use of relatively expensive high-memory EC2 instances. But because the job was architected to tolerate disruptions, researchers could specify the use of spot instances, which, at the time of the run, were priced at about one-twelfth the price of the equivalent on-demand instance. Ultimately, the institute paid just less than $20,000 to for the million-plus core hours of processing.

The parallel universe of HPC cloud workloads
Despite the enormous promise of cloud and HPC, not every HPC job is a fit for the model -- far from it. Using cloud for HPC workloads is still a new, albeit growing phenomenon. It's being used for a small subset of HPC jobs -- notably, the easiest ones, said Steve Conway, IDC research vice president for HPC.

By and large, HPC workloads in the cloud tend to be what the industry considers embarrassingly parallel -- large jobs that can be easily split into smaller ones, because they have few or no dependencies between them. It is therefore possible to run the jobs simultaneously on commodity hardware nodes that are not connected by high-throughput or low-latency pipes -- the kind of low-end infrastructure that by and large makes up the public cloud.

But most types of HPC jobs outside that cloud subset require communication between tasks. For example, to report the results of an intermediate step, and thus are usually run on systems that feature high-speed communication such as InfiniBand fabric between the processing nodes.

Thus, is largely true that the cloud is the best fit for jobs with a high degree of parallelism, conceded Cycle's Stowe, but "there's nothing embarrassing about it.... I like to think of them as pleasantly parallel," he said.

At the same time, cloud providers have responded by offering new compute resource types that can be a better fit for a broad range of HPC jobs. Amazon offers Cluster Compute, comprised of eight instances running on memory-rich Intel Sandy Bridge nodes. The Cluster Compute instances can also be located in Placement Groups that guarantee low latency, with 10 Gbps bandwidth between instances.

Likewise, Microsoft recently began offering so-called Big Compute nodes on its Azure public cloud. These come with eight instances and 60 GB of RAM, or 16 instances with 120 GB. Both run on Intel Sandy Bridge nodes with DDR memory, five 1 TB disks, 10 GbE for network and storage communication and 40Gb InfiniBand for internode communication.

If that still isn't enough, some traditional HPC vendors are also getting in on the cloud act, renting out access to optimized HPC hardware that provides better performance than typical public clouds. Examples include Penguin Computing's Penguin on Demand and Cyclone from SGI.

Part of SGI Cyclone's performance advantage comes from the fact that, unlike most public cloud offerings, it does not include a virtualization layer, said Franz Aman, SGI's chief marketing officer. "Having apps run natively on physical hardware gives you the performance you need," he said. Cyclone also provides Software as a Service (SaaS) access to specialized HPC software stacks such as NUMECA for computational fluid dynamics, or Gromacs for molecular dynamics.

These sorts of high-performance, integrated cloud offerings provide a stepping stone that can jumpstart a small company's use of HPC, or augment an organization's existing resources quite nicely.

Not all sunshine ahead for HPC in the cloud
Whilst some organizations have had great success doing HPC in the cloud, others maintain that it still has a long way to go. In fact, many HPC cloud services have proved to be disappointing, said one technical project manager at a life sciences startup, who is not authorized to talk to the press. After conducting an extensive pilot last year, she concluded that most mainstream public clouds are not a good fit for HPC, and that most traditional HPC software stacks don't integrate well with cloud resources.

She was disappointed, for example, by EC2's small instance sizes and the lack of InfiniBand connectivity. Further, "EC2 is too much like a data center—you need a sys admin with DevOps skills to get it to work, because scientists don't want to tinker with this stuff." And traditional HPC management software vendors have yet to nail cloud integration.

Internally, the firm uses cluster management software from Bright Computing, which she described as "very stable, robust and user friendly." However, an initial test of Bright Cluster Manager's cloud bursting feature, which is designed to let administrators launch new clusters in EC2—or burst existing clusters—didn't go well. "It's not plug-and-play yet," she said.

Then there's the issue of data privacy, said Dr. John Meyers, assistant professor and technology director at Boston University School of Medicine. The school currently manages several hundred terabytes of data, mainly in support of next-generation genetics and genomics research, plus cellular imaging, but that data is crunched exclusively in-house. "We're leery of the cloud because a lot of our data is identifiable patient data" regulated by laws like HIPAA [the Health Insurance Portability and Accountability Act], said Meyers.

However, if and when the school does resolve its cloud privacy and security concerns, it is in a good position, he said, thanks to the school's use of storage virtualization software from Actifio that manages data copies for backup, disaster recovery, and test and dev purposes.

"If we do go down that road, the fact that Actifio is application-sensitive will make it easy for us to say 'Shoot this data off to the cloud, but keep this data in-house,'" he said. Other storage management platforms that operate at the lower block level don't make it as easy to segment data by type.

Moving data to and from the cloud is another frequently cited HPC challenge, said IDC's Conway. The time it takes to transfer data to the cloud depends on the size of the organization's Internet connection, and, of course, the data set. Even with a low-end 1.5 Mb/sec Internet connection, simple math says it should only take 84 minutes to upload a 1 GB file. But a 1 TB data set would take over 58 days to upload at that speed, and a more reasonable—but still unwieldy—21 hours on a super-fast 100 Mb connection. "Getting the data to the cloud for the first time can stop people cold," Conway noted.

Some specialized cloud providers offer data upload services, and it is possible to speed up data transfers by sending files in multiple streams and by using compression. And of course, some organizations have their data co-located at or near the cloud provider's site, rendering these issues moot. However, for everyone else, it is imperative to think about the time it will take to get data to (and from) the cloud before jumping in. The waters may be too cold for some people's taste, but for a few hardy souls, it's just fine.

 

The skinny on Mobile Application Delivery

Determining how to deliver mobile applications all depends on your environment and your users' needs. Each mobile application delivery option comes with its downsides. Once you've decided to support mobile applications, you have to figure out how to deliver them. You could use device app stores, the cloud, the Web, virtualization, an enterprise app store -- or a mix. Keep in mind, however, that you should assess the costs of mobile application delivery and how to deal with getting apps to users without Internet connections.

How can I deliver applications to mobile devices?
There are a few different app delivery options you can mix and match to meet your mobile application delivery needs. Enterprise app stores are a good option because they give IT control over data governance, volume purchasing and licensing, but homegrown app stores are tricky to implement. You can use Web apps instead of native applications, which give you the advantage of device agnosticism but can limit app functionality.

Then there's desktop virtualization for mobile devices, which lets you deliver a PC environment to a smartphone or tablet. This approach keeps sensitive information in the data center where IT can control it, but it's not often an enjoyable user experience. Using the cloud is another option, but you have to pay for an existing service, such as Dropbox, or build a cloud storage service -- both of which can be expensive.

How can I use the cloud to deliver applications?
There are three ways to use the cloud to deliver mobile apps. You can build an internal cloud, use an external one, or do both. The disadvantage of an external cloud is that the applications you're trying to deliver are posted to that cloud's server, but an external cloud could work if your apps can be public. To keep apps private, it's possible to post them on a secure Web server that employees can access to download apps wirelessly. For this mobile application delivery option, you'll need to build your own cloud.

Can I use application virtualization to deliver apps?
Mobile app virtualization is one way to deliver apps to employees. It's faster than rebuilding existing apps for a mobile environment. For example, you can virtualize an existing Windows 7 application and send it to users' mobile devices.

Still, just because you can virtualize an app doesn't mean you should. Often, accessing a virtual application on a mobile device isn't a great user experience. If mobile versions of applications your organization uses are available from device app stores, it might be better to let employees use those and virtualize only the apps they can't get their hands on. Also keep in mind that delivering virtual applications to mobile devices only makes sense cost-wise in organizations that have already virtualized other apps.

What are some potential obstacles when deploying mobile apps?
Internet connectivity is a big one. Depending on your mobile application delivery method, users need to be connected to the Internet to access those apps and data. Also think about cost. You not only need to fund deployment, which can be through an enterprise app store, self-service portal or the cloud, but you also need to pay for application development and lifecycle management. All of those costs can significantly affect how -- and even if -- you deploy mobile applications.

 

Best practices for building your company’s next mobile application

What’s your company’s next mobile app project?
Along with the upswing in Web application development, many businesses are also trying their hand at creating customized mobile apps. These projects can involve creating a mobile version of a Web application or building a completely new app specifically for use on mobile devices.

Why build instead of buy?
In some cases, it’s possible to use off-the-shelf apps for business. That’s often the case with file sharing software for small to mid-sized businesses since file formats are fairly standardized across industries. If you can share MS Office documents and spreadsheets and PDFs, you’ve pretty much got things covered with popular apps like GoogleDocs or DropBox. However, there are many businesses that do have processes which require bespoke mobile apps. This is most likely to be the case for:

• Large enterprises (needs tend to be more complex)

• Niche industries (needs tend to vary from the average)

• Apps that are linked to customized internal databases

• Apps that must be integrated with an ERP

• Apps that require greater than average security and control

What types of apps are most popular?
65% of respondents to a 2011 Survey put business apps at the top of the list for the type of mobile application they will be developing in-house. Another 20% list finance apps as their main focus. Of course, the term “business” isn’t all that specific. That’s partly because there’s really no limit to what an imaginative organization can create to make the most of its resources in the new, mobile era. Here are a few bespoke mobile apps that could add value for businesses:

Crossover financial transaction apps
The custom-built business finance apps of the future are likely to be multifunctional. This goes way beyond building a mobile cash register. There are many mobile apps already developed that allow sales reps to accept credit card information and generate electronic receipts in the field. Some can even be interfaced with accounting databases and software like QuickBooks that many companies are already using for basic A/R purposes. 

However, a more advanced version might also offer customer self-service and be able to pull relevant data to enhance the consumer experience. This would mean integration with the customer’s past order history, the ability to handle returns and other non-standard transaction details, and perhaps even targeted marketing with coupon codes.

Other mobile finance apps might allow users to track expenses and make certain categories of purchases on behalf of the company based on predetermined parameters or interactive authorization with mobile users up the “approval path”. This type of application might be useful in logistics. Truckers could have access to funds for emergency repairs and track fuel purchases along with mileage using their mobile device. Ideally, this type of mobile app would be interfaced with the company’s logistics tracking software that keeps GPS tabs on vehicles at all times and logs trip information automatically.

Mobile ERP applications
Any app that ties into the core operations of a business is likely to be an attractive option for mid to enterprise level firms. Mobile doesn’t have to 

mean users are far-flung geographically. Sometimes, mobile technology can be just as useful on-site as off-site. For example, barcoding and inventory management points of contact that are currently managed with stationary equipment might just as easily be handled with mobile devices. This actually sets the stage for strategic expansion. Rolling out a mobile app for use in production through distribution and even retail might prove simpler than attempting to standardize an upgraded electronic inventory management system throughout a supply chain. The issue, of course, is designing a custom app that is compatible with a broad range of mobile OS and hardware. Otherwise, you face the challenge of standardizing mobile devices across the supply chain.

Customer focused mobile apps
This is the area where customization really goes wild since it involves the greatest opportunities for branding and interactivity. However, smart businesses will still focus on what they’re actually achieving with mobile apps rather than how flashy the experience is. Driving revenue means keeping the UI simple and giving customers what they actually want. Take industries such as direct sales where customers are expected to fill out lots of forms to get a quote or make a purchase. A mobile app that auto-populates fields from one form to the next is a real time saver and a powerful sales tool.

A somewhat riskier approach that can still pay off is the creation of business apps solely for the purpose of building brand loyalty. This might include:

• Loyalty program apps that let customers build and redeem points

• Entertainment apps (games) that create a buzz about the company

• Educational resources (free information that builds credibility and trust)

These are the types of business apps that can “go viral” or simply be a flash in the pan. One way to choose concepts that catch on is to make sure that 

sufficient business intelligence is collected beforehand. This is actually an area where other mobile apps (especially those that interface with social media) can shine. We’re likely to see a rapid expansion in the type and functionality of mobile apps for business over the next few years as enterprises cash in on the customer data they’ve been accumulating over the past decade.

 

 

Avoiding Cloud provider lock-in

This addresses one of the major concerns users have had about putting too much of their data -- or more importantly, their most valuable or critical data -- into the cloud. Put another way, how can users gain cloud storage benefits such as ease of accessibility and reduced costs, without the fear of vendor lock-in?

This issue is on the minds of many end users we speak with, and unfortunately, there's no simple or universal answer today. Fear of vendor lock-in is clearly a legitimate concern. There are no widely adopted standards in place today to ensure that customer data can be freely moved among different cloud storage service providers’ sites. And it always seems to be more difficult and more costly to transfer data out of a cloud storage repository than it is to upload it in the first place. The more data a user has, the harder it is to move; and if the user wants to move the data to a new cloud provider, they’ll probably have to pay for the bandwidth twice. This inhibits opportunistic migration and puts customers in a weak negotiating position relative to their cloud storage service provider.

So what steps can cloud storage customers take to reduce the likelihood of lock-in, and the cost and inconvenience that go along with it? Here are a few guidelines that we recommend users follow as they shop around for a cloud storage provider:

• Read the fine print of each provider’s policies, and if necessary, ask them directly how they facilitate moving customer data out of their cloud storage repository. Given the amount of data you’ll be uploading and its expected growth over time, could the data be moved via the Internet back into your data center in a reasonable timeframe or to a different provider’s site? As an alternative, if the volume of data is too large for digital transfer, can it be moved via a portable storage device? What's the process, timeframe and cost required for each of these approaches?
• Unfortunately, today, pulling data out of most cloud storage solutions requires a brute-force approach that you, the customer, will be responsible for. But it’s worth finding out the likely scenarios before you sign up with a provider. You may not like all the answers, but at least you’ll have a feel for the magnitude of effort and cost you might incur down the road.
• Ask the provider whether they offer data migration tools or services to facilitate the movement of large amounts of data. For example, while most providers require that customers moving data between clouds first download data to an intermediate location (e.g., the customer’s data center) and then re-upload the data to a new cloud, Nirvanix provides direct cloud-to-cloud data migration via its Cloud Sideloader technology. Today, this service is available only to move data into a Nirvanix cloud, but customers don't need to pay inbound bandwidth charges. Some cloud gateway vendors are also making migration easier by integrating with different cloud storage APIs and then providing a standard file system interface to facilitate data migration between those clouds.
• Choose providers that have pledged to support emerging industry standards, such as the Cloud Data Management Interface (CDMI) standard created by the Storage Networking Industry Association (SNIA). The CDMI provides a standard functional interface that applications will use to create, retrieve, update and delete data elements in the cloud and, once adopted, will make it much easier to move data from one cloud to another. The OpenStack initiative is also attempting to create and enforce standards that will facilitate data movement across different platforms and providers.

As the market matures, we’re hopeful that cloud storage service providers and technology vendors will rally around a handful of cloud standards that will put the data and vendor lock-in problem to rest. But in the meantime, let the buyer beware.

 

Android fragmentation complicates enterprise BYOD support

Samsung has made strides in making Android enterprise-ready, but the operating system remains too fragmented for most organizations to standardize on.

There are six different versions of the Android OS currently in use. The latest statistics from Google this week showed that 94% of the nearly 500 million Android devices activated are powered by Gingerbread 2.3 or higher. Ice Cream Sandwich 4.0 accounts for 29.3% of devices, Jelly Bean 4.1/4.2 accounts for 25% of devices and Gingerbread 2.3 is found on 39.8% of devices. Other versions of the OS run on about 6% of devices.

This level of fragmentation presents challenges to IT departments' mobile strategies because they need consistency across the Android platform. They are "fighting an uphill battle because they don't have much control over what devices get brought into their environment," said Benjamin Robbins, principal at Palador Inc., an enterprise mobility consulting firm based in Seattle.

Companies with a bring-your-own-device (BYOD) policy have an easier time limiting the variables they have to control and secure with Apple iOS, Microsoft's Windows Phone or BlackBerry than they do with Android, said Adam Bookman, co-founder at Propelics Inc., a mobile application development and consulting shop in San Jose, Calif. "With Android those choices are exponential," Bookman said.

The seemingly limitless choice of Android devices for consumers requires IT pros to make choices of their own when it comes to supporting the platform in a BYOD environment, Robbins said. Since most Android devices run version 2.3 or higher, IT can use that as a baseline to set a BYOD policy around. Employees who want access to the network and mobile management resources must have at least Android 2.3 installed on their device, Robbins said.

"That would certainly make life easier for IT in the short-term as they figure out how to support a range of platforms," he added. "It's better to take a device-agnostic mobile approach and concentrate on securing the apps and data because fragmentation isn't going away over the long term."

Android limitations lead to fragmentation
The Android Software Development Kit (SDK) has increased from 72 MB to 203 MB from versions 2.0 to 4.2, respectively. The inability to physically resize the partition containing the OS is one of the biggest reasons why fragmentation exists in the Android platform, said Lori Sylvia, executive vice president of marketing at Red Bend Software, a Waltham, Mass.-based company.

Red Bend's vRapidMobile is the technology behind firmware-over-the-air (FOTA) updates for many mobile device OEMs. The latest version allows OEMs to resize partitions over the air by taking extra space needed for the new OS version from the cache partition. The new feature does so without touching the space designated for the user's apps and data, although Sylvia could foresee Android OEMs allowing that to be a user option. The ability to resize device partitions won't work for Android devices bought within the last few years, but vRapidMobile will support FOTA updates for future devices using the technology.

Red Bend hopes that by eliminating the partition size limitation of OS upgrades more OEMs and carriers will be inclined to support devices for a longer duration, Sylvia said. However, even if the majority of OEMs push out timely firmware updates so that a majority of devices run the same OS version -- something IT pros would embrace -- there will always be form-factor and screen-size fragmentation to consider, Propelics' Bookman pointed out. Further, the OEM customizations of Android aren't going away because that's the primary way Samsung, HTC and other OEMs distinguish themselves from one another, he added.

Google, for its part, has made several attempts to address the problem of fragmentation. It launched the Android Update Alliance in 2011 as a joint agreement with OEM partners to support OS updates for at least 18 months after the release of a device. It also made Android's source code available to those partners earlier in the development process so OS updates would happen sooner, launched Google-branded hardware, and adjusted the Android SDK Terms of Service in January 2013 to limit OS forking.