Traditionally, software architects and developers would have expected their applications to run on a powerful server. Recently, however, the trend towards horizontal scaling has been increasing. Rather than expecting applications to run on highly scalable servers, developers havebeen redesigning (or refactoring) Information Systems and software applications so that they can scale horiizontally across a number of computer servers. This refactoring of applications is not a trivial task as both the applications and the data captured, managed and stored by these applications must be designed so that both processing and data can be broken down into smaller chunks. It is this existing architectural trend that has been a key factor propelling the adoption of cloud computing.

There are examples of horizontal scaling in High Performance Computing (HPC), Database Management Systems, CPU-intensive processing and Data-intensive processing.  Horizontal scaling had been used for HPC workloads long before the advent of cloud computing in a Grid Computing framework. Developers have refactored applications to achieve the distribution of HPC workloads across bare-metal compute grids. HPC has been used in many scientific applications. For example, scientists have broken down data for applications such as 3-D climate modelling so that it can be spread across a large number of servers. Grid computing is a predecessor to cloud computing as it uses tools to provision and manage multiple racks of physical servers so that they can all work together to solve a problem. As HPC is extremely demanding in terms of compute power, interprocess communication and input-output (I/O), such workloads would be most suitable for Clouds that provide Infrastructure As A Service (IaaS). Access to bare-metal servers or Type I Virtual Machines (VM) that provide more direct access to I/O devices would be specific examples.

As a sidebar we will define the terms bare-metal and Type I VMs. Bare-metal refers to the underlying physical architecture of a computer or server. Running an Operating System on bare-metal refers to running an unmodified version of the OS on the physical hardware. A Type I or Native VM refers to the scenario where the software layer that provides the virtualization for the VM runs on the bare hardware. Given that many HPC applications leverage the hardware directly for purposes of speed, it is clear that Bare-metal servers or Type I VMs would be suited for such applications.

Database Management Systems can also be adapted to run in cloud computing environments. Database servers can be horizontally scaled and database tables can be partitioned across the servers. This technique is known as sharding and allows multiple instances of database software, be it Oracle, MySQL, SQL Server or any other type of database, to scale performance in a cloud computing environment. Rather than accessing a single, central database, applications now access one of the many database instances depending on which shard contains the requested data.

CPU intensive applications are also good candidates for horizontal scaling. Applications that perform intensive tasks such as frame rendering (the process of transforming logical objects such as points, lines etc. into physical representations) can create a separate VM to render each frame rather than creating a new programming thread, thus enchancing performance. Horizontal scaling is also suitable for data-intensive processing as large amounts of data can be processed and the results coalesced to a coordinating process. For example, Hadoop (http://hadoop.apache.org/), which we discussed in a previous blog, is an open source implementation of the MapReduce framework for processing huge datasets using multiple computers.

The question we will ask in this blog is the role Horizontal Scaling can play for Smart Objects and Smart Networks. As we have illustrated in previous blogs, smart objects provide a rich new pool of real-time or near real-time data that will require processing. This data will need to be processed and stored. Frequently, the data will be need to be converted to meaningful information. For example, applications processing Wireless Sensor Network (WSN) data may have to apply complex mathematical formulae to convert measurements from engineering units to a more meaningful metric. When multiple instances of such data is arriving to a system every second it is clear that such applications may be both CPU and data intensive and would benefit from a horizontal scaling strategy. Similar logic also applies for storing smart object data. It may be difficult to predict the data storage requirements for smart networks at the outset. Rather than running the risk of a single database server becoming full, the data can be categorised and distributed across the cloud.

The use of data from smart networks for intensive tasks such as data mining, prediction formulation and pattern detection/analysis would certainly be CPU and data intensive tasks that would also be candidates for a HPC application. HPC could also be used for simulating and testing smart networks. One example is the use of HPC by the US Army Redstone Technical Test Center to test Ad-Hoc Wireless Sensor Networks. Given the role of smart objects in environmental monitoring and scientific applications - for example, biosensors to detect the presence of chemicals or other agents or monitor human and animal health - one would expect the use of HPC to process smart object data to grow in the coming years.

The Cloud Computing Infrastructure Models that are commonly defined are public, private and hybrid clouds. Each of these infrastructure models entail trade offs. It should be noted that the terms public, private and hybrid do not dictate location. Public Clouds are typically available over the Internet and private clouds are usually located at the premises of an organisation. However, private clouds could reside at a co-location facility as well and accessed over the Internet.

There are a number of considerations for organisations with regard to the cloud computing model they select. Large organisations with comprehensive and complex IT needs may need to avail of more than one model to solve different problems. For example, an application that is only needed on a temporary basis might be best suited for deployment in a public cloud as this avoids the need to purchase additional equipment that will only be used temporarily. Similarly, it might be best to deploy an application that will be used permanently by the organisation in a private or hybrid cloud. This also applies where there are particular requirements regarding quality of service or the location of data.

Public Clouds are run by third parties and applications from different customers are likely to be mixed together on the cloud's computer servers, storage systems and networks. It is typical for public clouds to be hosted away from customer premises. The key value of public clouds is that they provide a way to reduce customer risk and cost by providing a temporary extension to enterprise infrastructure. As stated in Sun Microsystem's white paper on Cloud Infrastructure, the existence of other applications running in a public cloud should be transparent to both cloud architects and users if a public cloud is implemented with performance, security and data locality in mind.

One of the key benefits of public clouds is that they can be much larger than most organisations' private clouds could ever be. Public clouds offer the ability to scale up and down on demand while the risks associated with deploying and maintaining IT infrastructure is shifted from the enterprise to the cloud provider.

Portions of a public cloud can be carved out for the exclusive use of a single client, thus creating a virtual private data centre. Rather than being limited to deploying virtual machine (VM) images (i.e. a software implementation of a computer that executes programs like a physical computer), a virtual private data centre gives customers greater access and control over the IT infrastructure they're using. Customers can manipulate not just VM images but also servers, storage systems, network devices and network topology. Creating a virtual private data centre with all components located in the same facility lessens the issue of data locality and overcomes potential bandwidth bottlenecks as all resources are in the same physical location.

Private Clouds are built for the exclusive use of one client and provide that client with a high level of control over data, security and quality of service. The customer owns the infrastructure and has control over how applications are deployed on it. Private clouds may be deployed at the organisation's premises in, for example, an enterprise data centre or may be deployed at a co-location facility. An organisation's own IT department can build and manage their own private cloud or a cloud provider can perform this service. Sun Microsystems refer to this latter service as a 'hosted private' model where IT infrastructure is installed, configured and operated to support a private cloud for the organisation. The key value of the 'hosted private' model is the high level of control it gives organisations over the use of cloud resources while bringing in the expertise required to establish and operate the environment. 

The final type of cloud infrastructure model is the hybrid cloud. This combines both the public and private cloud models and provides scalability from a third party cloud provider as and when an organisation requires it. The ability to extend a private cloud with the resources of a public cloud can be used to maintain service levels when workloads and data processing.storage requirements increase. Sun Microsystems cite the use of storage clouds (i.e. storing data over the cloud) to support Web 2.0 applications as a common example of the use of hybrid clouds.

A hybrid cloud can also be used to cope with planned increases in workload. This is sometimes referred to as 'surge computing' where a public cloud can be used to perform periodic task that can be deployed easily on a public cloud. However, it should be noted that hybrid clouds do complicate cloud architecture for an organisation as they introduce the challenge of deciding how to distribute applications across both a public and private cloud. The relationship between data and processing resources must be considered. If large amounts of data must be loaded into a public cloud for a small amount of processing one should question the efficacy of using the hybrid cloud model. As a general rule of thumb, a hybrid cloud will be more effective if the amount of data being transferred is small or the application is stateless (i.e. doesn't have to maintain setting and configuration data).

The question we will now consider is what implications the different Cloud Infrastructure Models have for smart objects. The volume of data produced by smart objects will greatly expand the data processing undertaken by organisations in coming years. Will smart objects and smart networks be considered just another service offered by cloud computing? Probably not, as smart objects and networks will be deployed locally. By their nature, smart ecosystems provide computing functionality and devices for infrastructure such as buildings, pipelines and natural environments. While these systems may be deployed remotely from an organisation's premise this is not cloud computing per se. However, the different cloud computing infrastructure models do offer different benefits to these smart ecosystems.

As noted, public clouds offer computational facilities and IT infrastructure that would otherwise be unavailable to the typical enterprise. Given the requirements to capture and interpret data from potentially millions of smart objects, the data processing required would be ideally suited for a public cloud. This data processing could include translating the data into a meaningful measurement, interpreting the data, performing pattern analysis, data mining and generating Business Intelligence. Such computations are intensive and will often require the computing power offered by public clouds.

However, one should not disregard private clouds for smart networks. In the case of private clouds located on premises, the transfer of data from smart networks (such as a smart building) deployed locally will require much less bandwidth than transferring this data to a public cloud. Often, the smart network will be part of the overall network used by the organisation for its private cloud. Private clouds provided remotely will also have potential benefits for smart networks  as the use of dedicated resources for processing and storing smart network data provides organisations with a finer level of control for managing the data provided by smart ecosystems.

Hybrid clouds and the affiliated strategy of surge computing may also be appropriate for smart networks in certain cases. The ad-hoc nature of smart networks implemented using wireless sensor networks, Bluetooth, Zigbee and other technologies means that demand for data processing and workload requirements can vary greatly. In such a circumstance, the ability to expand a private cloud with public cloud resources could be beneficial. However, the caveats regarding data transfer that were noted earlier should be taken into account.

There is no definitive answer as to the type of Cloud Computing Infrastructure Model that should be used for smart objects. It depends on the location of the smart network and the type of data processing being carried out. For smart meters, for example, the data is remotely captured at the utility customer's home or premises so one could argue that it is a moot point whether data is transferred to a (local) private or public cloud as the cost of data transfer may roughly be the same. This may apply to remote wireless sensor networks as well. However, there are other considerations regarding security and privacy that may mandate the use of a private cloud. In essence, organisations should consider the benefits and shortcomings of each Cloud Infrastructure Model for their smart networks and factor in legal and business as well as technical requirements.