Cloud Computing and the Future of the Enterprise Data Center
The enterprise computing world is seeing a revolution last seen during the evolution and adoption of the internet
Dec. 2, 2010 10:15 AM
This article discusses how the enterprise data center is dissolving into a virtual computing space where all computing resources are addressed through the Internet. It discusses the evolution of cloud computing, its affect on enterprise IT strategy and the future enterprise computing in the data centers as we know it.
The enterprise computing world is seeing a revolution last seen during the evolution and adoption of the internet. The brick and mortar Data Center, centric to all the computing needs of an enterprise is under pressure as CIOs crunch ROI and TCO numbers. The ever increasing cost of maintenance of an Enterprise Data Center is forcing business to look at these numbers and the implementing strategic decisions to cut cost. The common refrain is the ever increasing cost of maintaining a large Data Center, heating, cooling, electricity and maintenance costs, both hardware and resources.
This ROI and cost approach has been historically focused on the hardware and the facilities.
With ever-increasing faster and greater computing needs, business computing went with bigger machines, multi-processor, and multithread servers that could handle large application loads. These machines needed a large initial investment and an incremental cost in maintenance support. In most cases the capacity utilization of these machines was low, leading to a poor ROI. IT managers then moved their application to smaller footprint machines, and the blade system with multiple blades in the same chassis sharing a small footprint have become increasingly popular. The modular nature of the hardware and the ability to hot swap components has made this move very popular with IT managers.
The next step in the evolution to smaller footprint computing as well as increased utilization of the hardware was the adoption of virtualization strategies across all platforms. With the advent of virtualization in almost all areas from Wintel, to Unix to storage, a large number of virtual machines could be hosted on the same physical machine and multiple applications and users serviced from these virtual machines. This has meant that the system administrator does not have to be in close proximity to the server; he just needs to be able to access the machine over the network.
With the advent of virtualization as well as the large push to outsource and offshore application support as well as the easy and cheap access to relatively large bandwidths becoming available, moving the support of these virtualized machines to either outsourced service partners or even off shored partners has become a reality. More and more enterprises who have had long standing fruitful vendor relationships with their outsourced or off-shored partner are making a concentrated move towards remote management services of their enterprise hardware.
In all of the cases above the hardware remains to be hosted out of a brick and mortar Data Center. Large enterprises who own and manage their own data centers continue to bear the cost of operating the facility, the cost of the real estate, power, cooling and maintenance and upkeep of the back up systems, as well as the resources needed to operate the data center. The smaller enterprises have moved to co-located data center facilities that manage the facilities for a cost. All of them own and support their own hardware or at least pay for it through outsourced partners.
With advent of cloud computing and the successes of various IT functions that service their users with Applications and platforms over the internet, it is becoming clear that this method of IT service and delivery will fundamentally change the future of the Data Center relative to an enterprise. Cloud computing merges many new technologies into an amalgamated whole including web 2.0, fast and cheap bandwidth, virtualization, utility computing etc.
The North American National Institute for Standard and Technology (NIST) defines cloud computing as " a model for enabling convenient, on demand access to a shared pool of configurable computing resources that can be rapidly provisioned" and released with minimum management.
The environment in cloud computing can be examined in many different ways; for all intents and purposes these can be divided into three major area; Infrastructure As a service, Software as a service and Platform as a service. Each of these main areas can be divided into smaller fine grained areas of interest depending on what the enterprise is interested in moving to the cloud. Some pundits have had more categories( Storage as a Service, Database as a Service, Information as a Service, Process as a Service, Application as a Service, Platform as a service, Integration as a Service, Security as a service, Management as a Service, Testing as a Service, Infrastructure as a service).
Thus all aspects of an enterprise IT can be serviced out of a cloud.
The cloud computing environment has also been divided into Private Clouds, Public Clouds, Hybrid Clouds
Public cloud or external cloud describes cloud computing in the traditional mainstream sense, whereby resources are dynamically provisioned on a fine-grained, self-service basis over the Internet, via web applications or web services from an off-site third-party provider who shares resources across multiple client enterprises and bills on a fine-grained use based utility computing basis.
Private cloud and internal cloud is a phrase that are used to describe offerings that emulate cloud computing on private networks. These products claim to "deliver some benefits of cloud computing without the pitfalls", capitalising on data security, corporate governance, and reliability concerns.
A hybrid cloud as the name suggests consists of multiple internal and/or external providers delivering applications to the desktop through a ‘thin client' connection either over the internet or intranet.
Moving to the Cloud
The core steps in moving to the cloud can be divided into four successive steps.
- Analysis and Test
- SP Selection
- Migration and Support
After a strategic decision has been made by the enterprise to move to the Cloud, the most important challenge to the enterprise is to identify what can be moved to the cloud and in what order. This includes applications, services, data storage and retrieval etc. The maturity of processes in the enterprise as well as its state of virtualization will be important in making these decisions.
Many enterprises may want to test the cloud computing environment by moving their ‘non-core' applications to the cloud. This implies that the enterprise has a very mature BCP and DR process in place so that suitable ‘non-core' applications can be identified. This holds true for Enterprise data, storage etc.
Application integration technologies and methods also impact the choice of applications to be moved to the cloud. A highly complex and integrated application and set of services may be a good candidate for a private cloud, where as stand alone applications may be good candidates for a public cloud. Same holds true for enterprise data; confidential, trade related data may be hosted in a private cloud to manage security concerns and non critical data may be hosted in a public cloud. Hence cloud types and configurations could be managed to the risk and security profile of the enterprise.
Analysis and Test
Once the applications or platforms have been selected, it is essential to test these under the cloud scenario. If this will be accessed through a private cloud, a testing sandbox that comprises of the application, storage, intranet access etc needs to be created for testing. This is a validation exercise to test performance under various and load conditions. The results of this testing should align with established SLAs with the business, else moving to the cloud can adversely affect business functions.
Once the data, process and service of the identified candidate applications/process/DB has been completed, the service provider selection should be based on those needs. In addition business requirement on the viability of the SP, their ‘always-on' ability, the use of tracking/preventing machine downtimes; recover from failures dynamically, and the SLAs between the enterprise and the SP need to be considered while choosing the right SP or SPs. Depending on the criticality of the Application or process and the BCP associated with it, these may be hosted by multiple SP with failover capabilities between the two sites.
This is the last step in the process where application code, storage etc is ported/migrated to the cloud. This could involve the creation of new services and processes to support the new environment. This is an incremental process as the enterprise IT gets comfortable with the migration process and applications can be moved to the cloud in phases or batches.
The maturity of the enterprise computing is very important as companies look at moving to the cloud. A very mature BCP and DR process as well as highly virtualized Data Center environment will be a pre-requisite before companies can make the jump. The definition of core and non-core applications and data and the choice of the initial applications to move to the cloud is also critical. Some legacy graphical applications which are multi-desktop at the user end are not good candidates to move to the cloud. Applications whose performance has been fine tuned to the extent that any change in hardware environment creates adversely affects the performance are not good initial candidates.
As more and more companies start looking at cloud computing as a viable alternative to a brick and mortar data center, the future of the Data center as we know it is expected to change drastically. As risk concerns of moving applications and data to the cloud are removed by service providers getting the Cloud Data Center audited inline with COBIT, PCI-SOX and ISO27002 to prove that the data in their data centers is safe and various environments are physically and logically separated, more and more CIOs are expected to support their Data Center environments to the cloud (up to 65% per Gartner's report). Enterprise Data Centers, NOC, etc., may be dying breed in the next decade.