Cloud computing has been on the minds of senior IT and business leaders for quite some time now. Unfortunately, there has been a fair amount of confusion regarding cloud computing and its relationship with grid computing. Obviously, you cannot deploy something, let alone design it, unless you define it properly.
With cloud computing, large pools of systems are linked together to provide the necessary IT services required by the business. This enables the delivery of services from remote systems and networks (in the cloud), which share these resources (including network bandwidth) to any device from anywhere. This scalable computing model moves away from physical networks and servers in favor of virtual ones. These virtual-type resources are then provisioned and managed as transparently as possible to the client. The advent of Web 2.0 applications, along with the need to collaborate and network through the Internet, has made the need for cloud computing even more relevant.
How can cloud computing elevate technology to deliver services to lower Total Cost of Ownership (TCO) while increasing Return on Investment (ROI)? What does IBM® offer in the way of cloud computing and how do POWER systems fit into the picture? Why is cloud computing relevant to the client and how is cloud relevant to AIX® and the System p®? This article answers these questions. The article discusses the datacenter and how implementing cloud computing with Power systems can help datacenters act more like the Internet, deploying a Service Oriented Architecture (SOA) inherent in cloud computing.
First, you must clearly understand cloud computing. Cloud computing defines a model where specific services are assigned to systems that are accessed through a network. These systems are pooled together to provide you with the required IT services. With cloud computing, customers do not usually own the infrastructure; they rent it. This is important from a TCO perspective because it avoids a massive capital expenditure. How is this different from grid? Grid computing is a form of distributed computing where the centralized computer is composed of a cluster of networks. Its primary purpose is to provide a high-performance infrastructure to support a specific purpose, which is normally one specific application.
Cloud is similar to grid in that they both require software to access external components on other networks. Grid provides an infrastructure that you turn it on and off as needed. You pay for what you need. While grid provides the infrastructure, it's the cloud that actually provides the application and service. Cloud is on-demand and is service oriented. Cloud has actually evolved from grid, and organizations focused on service delivery are increasingly understanding the need to provide the application-delivery transparency that is inherent in cloud computing, but is not really a part of grid technology.
Why is cloud computing so important? It's important because of the inherent value it provides its clients in terms of reduced TCO and increased ROI. It provides for increased efficiencies and resiliency in the shape of reduced complexity towards systems management, more efficient workload utilization (by sharing resources across multiple applications), and optimum application deployment flexibility. It allows customers to develop, deploy, and run systems that are highly reliable, scalable, and perform extremely well. And it does all of this while removing the worry about where the actual infrastructure resides.
Some characteristics of cloud computing include self healing, virtualization, and its ability to support green computing. Cloud computing supports green computing with its ability to only use required resources. Green computing refers to the concept of increased energy efficiencies, usually in datacenters. Cloud computing complements green computing because of its on-demand focus. You are only using what you need to. The typical datacenter is built to support all of the infrastructure necessary to run the corporate applications of the company. Cloud computing is service oriented and is only utilized when that service is required.
Looking at the bigger picture, clouds can utilize grids (the underlying infrastructure), but not the reverse. The traditional datacenter is always on, waiting for the next processing or batch cycle. Cooling is needed constantly, as is energy to power all datacenter systems, whether they are running or not. Green datacenters typically are built around the idea of maximizing datacenter resources, using virtualization software as its underbelly. Power consumption is controlled by management software. Cloud computing resolves these issues.
In fact, IBM, with its new Enterprise Data Center strategy, fully supports cloud, green, and grid technologies. It does this by using the cloud model, which allows for the delivery of services to clients, regardless of available computing resources. It also does so in such a way as to enable delivery based on actual demand, which further cuts costs. In a way, it allows the corporate datacenter to operate more like the Internet, because computing is now distributed around a global fabric of resources, as opposed to local datacenters housing local server farms. It should also be noted that many companies today are already taking advantage of cloud computing, including Google, Yahoo, Amazon, and eBay. Cloud computing is definitely not vaporware. it is very real and being utilized successfully today.
The new IBM new initiative, the Enterprise Data Center, provides for the use of technologies such as Web 2.0 and cloud computing through its model of delivering seamless access to services and information. This new approach to the datacenter has the following elements:
- Virtualized resources
- Maximization of power and cooling resources
- Business-driven service management
- Business resiliency and security
- Information infrastructure
Virtual resources allow for quicker provisioning, increased flexibility, more efficient workload, and a tighter integration between physical resources such as servers, network and applications, and storage. It allows IT to be more responsive to the business without impacting performance or reliability. Green infrastructure allows you to align power and cooling needs with real business requirements. In the process, it reduces the number of systems and networks necessary to run applications. As a result, it reduces costs and improves efficiencies. This service model allows IT to simplify infrastructure while maximizing availability, scalability, and performance. It also increases overall operating efficiencies. Business resiliency and security delivers more secure systems, ensuring that "best practices" models are followed. Infrastructure based on information systems helps eliminate disparate and silo-oriented systems and networks. While doing so, it increases the ability to deliver information to clients, while also optimizing reliability and performance. IBM has released studies showing that by utilizing its model, customers have been able to reduce heat by up to 60%, reduce floor space by as much as 80%, reduce energy consumption by up to 40%, and increase availability by cutting back on outages by 58% (see Resources). IBM currently has 13 cloud computing centers worldwide.
Here is a list of some of the IBM cloud computing initiatives and investments:
- The IBM cloud initiative started kicking into high gear in November of 2007 with its Blue Cloud™ announcement. Blue Cloud actually evolved from the IBM Technology Adaption program, a system where IBM developers requested computing resources to test innovations. The initiative is based on open standards and open source software supported by IBM technology and services. The cloud itself is based on the IBM Almaden Research Center Cloud infrastructure, which includes IBM PowerVM™ virtualization. Blue Cloud offerings also includes POWER-based systems, which holds particular interest for AIX clients.
- On January 26, 2009, IBM announced its work with six universities, leveraging Blue Cloud to step up research initiatives, which were constrained by time, limited resources, and systems that were overloaded and overworked. Part of this announcement included the Qatar Computing Initiative, which will open up its cloud infrastructure to local business and industry. One example of this is that it will help in the exploration of oil and gas.
- In June of 2008, IBM announced Africa's first cloud computing center. IBM put 120 million dollars into this effort to help in the effort to modernize and build out infrastructure, support government services, banks, and telecom. This center showcases Web 2.0 technology, next-generation banking systems, and a service oriented architecture (SOA).
- In July of 2008, IBM started announcing plans to establish a new data center at its research facility the North Carolina Data Center - and to use it to sell cloud computing services to its clients. This was a $360 million initiative, which initially provided for 60,000 square feet. The facility will use virtualization to reduce energy consumption by running multiple applications on the same servers.
- In February of 2008, IBM announced plans to establish the first cloud computing center in China, in Wuxi, called the China Cloud. This was a shared initiative, as IBM will work with local government authorities in Wuxi as well as its business partners. The shared facility will provide for cloud computing by replacing traditional datacenters where each company owns and manages its own systems, hardware, and software.
This section discusses how Power Systems™ and AIX lend itself to cloud computing. First, the article looks at AIX V6.1.
- Workload partitions (WPARs). This new feature, available only in AIX 6.1, allows you to have fewer operating system images (kernels) on your POWER servers. Similar to Sun containers, it allows you to have multiple application or workload environments, residing in one kernel or operating system partition. Probably the easiest way to qualify WPARs is to refer to it as virtual operating system partitioning. This innovation really lends itself to cloud in a big way because it allows you to deploy applications and services far quicker than ever imagined, usually within minutes. While hypervisor-based logical partitioning helps consolidate and virtualize hardware within a box, it is operating system virtualization that really allows for a more granular, flexible approach to systems and workload management. The end result is a much more efficient use of resources. At the same time, it provides IT with the ability to react much more quickly to the needs of the business by being able to deploy new workloads quickly.
- Live Application Mobility. This is another important new feature of AIX V6.1, which allows clients to relocate entire workload partitions without shutting down the application or WPAR. No other midrange hardware vendor has this ability and it is this type of innovation on AIX V6.1 that really lends itself to the cloud. It also provides cloud administrators with the ability to improve performance by moving workloads from servers that are heavily utilized to more lightly utilized workloads. Furthermore, it helps increase energy savings by allowing administrators to move workloads around during non-peak periods, such as nights or weekends, effectively shutting down hardware that is not required during these periods. Unlike live-partition mobility, which is a feature of POWER6™, live application mobility is only available on AIX V6.1. It can also be used with older POWER hardware.
- Security enhancements. Security is another important aspect of cloud computing. AIX V6.1 provides many enhancements to security, including role-based access controls (RBAC), encrypted filesystems, trusted execution, and enhancement to AIX Security Expert, Trusted AIX, and Long Pass Phrase Support. RBAC provides for allowing non-root system administrators the ability to perform system administration tasks, which is important for overall datacenter/security transparency, another important component of cloud computing.
Perhaps the most important technical driver for cloud computing available on POWER servers is the virtualization features available for AIX in the form of the IBM Virtualization engine, PowerVM. Features and functionality that are particularly important in a cloud environment include:
- Hardware Virtualization. This allows for the virtualization of hardware components. For CPUs, it provides the functionality to share resources across multiple operating systems from shared processor pools. You can actually borrow processor power from idle computers to handle increased workload on other partitions from a single managed server. This is done by uncapping partitions.
- Micro-Partitioning. This feature provides the ability for a single CPU to support up to 10 host partitions, each sharing as little as 1/10th of a CPU. This really helps in the effort to supply an almost unlimited amount of resources to those in the cloud that require them. It also allows you to provision resources quickly, meeting the stringent demands and requirements of the business, further allowing IT to become an agile partner in the business.
- Virtual I/O Servers (VIOs). This specialty partition provides VIO resources to clients that are using VIO servers. This includes storage (Virtual SCSI) and networking (Shared Ethernet).
- LX86. This new feature allows you to move x86-compiled Linux® applications to POWER systems without having to migrate, compile, or otherwise port the application natively to POWER. This is important because it allows you to scale vertically on Linux on Power (LoP) systems, eliminating the need for horizontal server farms of hundreds or thousands of x86 systems. In doing so, former x86 Linux clients will now be able to take advantage of the RAS, performance, and resiliency elements of POWER, all of which lend themselves to cloud computing.
- Partition Mobility. This feature, available only on POWER6 servers, allows clients to move AIX (or Linux) partitions from one physical server to another, without application downtime. This provides for increased availability of systems, and is a great feature for planned outages.
An important benefit (sometimes overlooked) of using POWER systems that lends itself to cloud computing is Capacity on Demand (CoD), which is available for both AIX and Linux partitions running on POWER servers. This feature provides the flexibility and on-demand functionality that is vital to cloud computing. CoD allows you to activate CPU and RAM on an as-needed basis. Traditional Capacity Upgrade on Demand (CuoD) is a system that allows you to permanently use active CPU or RAM that has already been purchased. Trial Capacity on-demand allows you to evaluate the use of inactive processes or RAM at no charge.
Recently, new features and functionality have been added to CoD that lend themselves to the cloud in a big way. It does this by allowing clients to temporarily activate and deactivate CPUs or RAM to meet the peaks and valleys of workloads. Essentially, you can start and stop requests all day long and you are only billed for your usage, billed at the end of the quarter. This is using the On/Off Capacity on-demand feature, a feature only recently added to POWER systems. Another new feature is Utility Capacity on Demand, which borrows from the concept of utility computing. This allows you to deliver increased capacity from the shared processor pools. Capacity backup adds to availability and allows for a minimum set of active processes that can be activated in the event of an emergency. This feature is integrated with PowerHA™, which will be discussed next.
PowerHA provides clients with the availability needed in today's environments where downtime is simply not acceptable. As a matter of course, cloud computing must have strong availability mechanisms in place. PowerHA is used for both planned and unplanned outages and provides the mechanics for one server to failover to another in case of a hardware event. It also has an optional feature, PowerHA/XD, which provides clients with business continuity in case of a major disaster. What are you going to do if the cloud goes down in the location that houses your systems? This product provides the cloud with that important disaster-recovery safety net.
POWER servers today not only look like mainframes (for instance, the IBM 595p) but are increasingly evolving into systems that share their inherent stability. One way they have evolved in this fashion has been by borrowing from the legacy of RAS (Reliability, Availability, and Scalability) on the mainframe -- IBM System z®. RAS capabilities are essential for cloud computing because the cloud's resources must always be up and running. Below are some recent RAS features that have been implemented on the new POWER6 architecture:
- Processor retries. CPUs in IBM POWER6 can now allow failed instructions to continue. Essentially, it stores all the data in the chip and reverts back in case of a prior error. Using CoD, it can either search for spare processors or move them to unused processors.
- First Failure Data Capture (FFDF). This feature allows the service processor (SP) to isolate faults and log events that take place right up to the point of the actual failure. In the event of a processor failure, dynamic deallocations of processors take place for systems that have at least two cores. This type of self-healing feature is a necessity for systems that reside in the cloud.
- Hot-node Add and Replace. This new feature/functionality is limited to top-of-the-line IBM 595p. It allows you to add, deactivate, and repair entire CPUs without a reboot. This type of resiliency is a key component of cloud computing because it provides the datacenter with the flexibility it needs to add and fix resources dynamically.
Other RAS features include dynamic firmware maintenance and hot-swappable disk bays for servers and I/O drawers. These features also increase overall datacenter resiliency.
While cloud computing may not be the panacea that many have portrayed it as, it certainly has a much better chance of being a hit than grid computing. The primary reason for this is that cloud computing involves business services, not just the infrastructure component, which is the story with grid. Furthermore, IBM and other companies have also invested considerable resources and effort into ensuring that there really is a demand for cloud and that the capabilities to implement cloud computing are actually available. This article discussed how the concepts around cloud computing naturally lend themselves to AIX and Power Systems. Some of the drivers include: virtualization (PowerVM), RAS, availability offerings, as well as new feature/functionality of both AIX and POWER servers. Cloud computing appears to have a bright future ahead of it, and POWER servers driven by AIX (and Linux) stand to be a big part of that future.
Get information on the upcoming International Cloud Computing Conference and Expo.
IBM portal resources: This is the place to begin exploring IBM's vision of cloud computing.
The New Enterprise Data Center portal: Learn about IBM's new approach for efficient IT delivery.
Read IBM's press
release on cloud computing implementation services.
Read the 2008 Server OS Reliability Survey from the Institute of Advanced Professional Studies.
Virtualization with coLinux (M. Tim Jones, developerWorks, March 2007): Read an article that introduces virtualization and then looks at the approach taken by Cooperative Linux (coLinux).
New to AIX and UNIX?: Visit the "New to AIX and UNIX" page to learn more about AIX and UNIX.
AIX Wiki: A collaborative environment for technical information related to AIX.
Optimizing AIX 5L performance: Tuning network performance, Part 1 (Ken Milberg, developerWorks, November 2007): Read Part 1 of a three-part series on AIX networking, which focuses on the challenges of optimizing network performance.
For a three-part series on memory tuning on AIX, see Optimizing AIX 5L performance: Tuning your memory settings, Part 1 (Ken Milberg, developerWorks, June 2007).
Read the IBM whitepaper Improving Database Performance with AIX concurrent I/O.
Learn about AIX memory affinity support from the IBM System p and AIX InfoCenter.
Learn about IBM's Power Architecture: High-Performance Architecture with a History.
Read Power to the People; A history of chip making at IBM (developerWorks, December 2005) for coverage of IBM's power architecture.
CPU Monitoring and Tuning (Wayne Huang et al. developerWorks, March, 2002): Read this article to learn how standard AIX tools can help you determine CPU bottlenecks.
For a comprehensive guide about the performance monitoring and tuning tools that are provided with AIX 5L Version 5.3, see the IBM Redbook AIX 5L Practical Performance Tools and Tuning Guide.
Learn what features you can benefit from in AIX 5L Version 5.3, in AIX 5L Version 5.3: What's in it for you? (developerWorks, June 2005).
Operating System and Device Management from IBM provides users and system administrators with complete information that can affect your selection of options when performing such tasks as backing up and restoring the system, managing physical and logical storage, and sizing appropriate paging space.
The AIX 5L Differences Guide Version 5.3 Edition (developerWorks, December 2004) redbook focuses on the differences introduced in AIX 5L Version 5.3 when compared to AIX 5L Version 5.2.
The AIX and UNIX developerWorks zone provides a wealth of information relating to all aspects of AIX systems administration.
Open source: Visit the developerWorks Open source zone for extensive how-to information, tools, and project updates to help you develop with open source technologies, and use them with IBM products.
developerWorks technical events and webcasts: Stay current with developerWorks technical events and webcasts.
Podcasts: Tune in and catch up with IBM technical experts.
Get products and technologies
Participate in the AIX and UNIX forums:
- AIX Forum
- AIX Forum for developers
- Cluster Systems Management
- IBM Support Assistant Forum
- Performance Tools Forum
- Virtualization Forum
- More AIX and UNIX Forums
Ken Milberg is a technology writer and site expert for Techtarget.com and provides Linux technical information and support at Searchopensource.com. He is also a writer and technical editor for IBM Systems Magazine, Power Systems edition, and a frequent contributor of content for IBM developerWorks. He holds a bachelor's degree in computer and information science, as well as a master's degree in technology management from the University of Maryland University College. He is the founder and group leader of the N.Y. Metro POWER-AIX/Linux Users Group. Through the years, he has worked for both large and small organizations and has held diverse positions from CIO to senior AIX engineer. He is currently president and managing consultant for UNIX-Linux Solutions, is a PMI-certified Project Management Professional (PMP), an IBM Certified Advanced Technical Expert (CATE), and is also IBM SCon certified.