Problem solve Get help with specific problems with your technologies, process and projects.

Using metro storage clusters to manage data growth, add value to IaaS

Metro storage clusters -- geographically distributed, interconnected storage platforms -- can help IaaS providers manage data growth and add value.

Data growth is the largest data center infrastructure challenge for large enterprises, according to Gartner, and service providers are experiencing it at an even bigger scale. Requirements for ready access to all data types has been growing exponentially, increasing pressure on storage professionals and storage platform manufacturers to solve the data availability and reliability problem -- from the transport layer up to the application layer.

Cloud Provider Spotlight Infrastructure as a Service (IaaS) providers and some large organizations are solving this problem by creating metro storage clusters, sets of geographically distributed storage platforms providing ready access to users when needed. These clusters are capable of spanning across data centers in multiple geographies using advanced storage and Layer 3 technologies. Some examples of these technologies include storage replication, virtualized storage clustering at the hypervisor layer, and dedicated Fibre Channel and network links between data centers across a city, state or larger geography. The data can be accessed from virtually anywhere with high availability and reliability.

There are three primary value propositions to consider when making any technology decision, all of which are equally important to take full advantage of the technology being selected. I call this the values triangle of a technology decision, which can be used to examine the business, technological and operational values of a purchasing decision. Cloud providers can apply this model to evaluate metro storage cluster adoption.

Values triangle of a technology decision

Business value of the decision

The primary reason any business should implement a given technology is that it adds value in some way. For enterprise customers, metro storage clusters satisfy multiple business challenges, including disaster recovery and business continuity, and they help address security and compliance requirements.

Service providers may gain significant business value by implementing metro storage clusters as well. They can use the technology to create multilayer data storage packages for customers, depending on which metro storage cluster design they chose to deploy. Requirements would dictate the design, architecture and services to be offered. Cloud providers could offer every layer of service within the data storage package at a premium service charge. Consider Replication as a Service or Disaster Recovery as a Service, or multisite and stretched clustering for managed services customers.

Technological value of the decision

Depending on the specific deployment, metro storage clusters could significantly increase the operational value of data management and availability.

The technological value of any decision needs to be based on empirical data, and the best way to identify the value is to look at the requirements. Metro storage clusters can solve multiple complex technology problems, such as multisite high availability, increased data reliability, stretched compute and virtualized clusters -- all of which translate into highly available applications enabling users to perform their tasks whenever necessary. User data is made available to compute clusters across multiple data centers, to the applications and ultimately to the users by means of inter-data center replication, storage area network routing, virtual SANs and other advanced storage and compute technologies.

What matters most to businesses today is data. That data can be in the form of an on-demand video service pushing gigabytes of high-definition streaming video across the Internet so the consumer can enjoy a movie night, or it can mean allowing an avid investor to make an instant trade. In either case, the data needs to be available to the application, and metro storage clusters enable this functionality and bring significant technological value to the business.

Operational value of the decision

The cost of overlooking operational value of any technology decision can be significant, both in terms of customer satisfaction as well as operational expense. Any technology being selected to solve a business and technological problem needs to conform to the basic processes, procedures and standards for business and platform operations. The ability to continuously and proactively monitor the platform and carry out simplified capacity management, performance management, configuration management and end-to-end change management are just some of rudimentary requirements for operational excellence that must not be overlooked when making a technology decision.

Depending on the specific deployment, metro storage clusters could significantly increase the operational value of the data management and availability. In fact, metro storage clusters provide the ability to spread data across multiple data centers on highly available, distributed platforms. Operationally, metro storage clusters enable service providers to offer higher uptimes to their customers by facilitating data availability across multiple sites.

Platform lifecycle management becomes simpler: Scheduled outages are not a concern, and change management can support scheduled changes at any time, as long as only a single site is affected and availability elsewhere is maintained. Platform administrators can simplify performance management using metro storage clusters by implementing such advanced functionalities as Quality of Service or read- and write-anywhere algorithms that streamline bandwidth consumption across available sites while managing data replication on the back end.

As long as decision makers take into account this values triangle using facts and empirical data, the selected technology will deliver substantial value to the bottom line, including the abilities to sell additional services on top of simple IaaS, to extract significant technological efficiency out of the deployment and to create a holistic, streamlined operational structure around the technology.

Said SyedAbout the author:
Said Syed is a manager of platform engineering and architecture at Verizon Terremark, a Miami-based business unit of Verizon Communications specializing in infrastructure and cloud services.

This was last published in June 2013

Dig Deeper on Cloud and Managed Network Services

PRO+

Content

Find more PRO+ content and other member only offers, here.

Join the conversation

1 comment

Send me notifications when other members comment.

By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy

Please create a username to comment.

Interesting article. Using an example of metro storage clusters in an application and/or database environment would have been helpful. Is it for automatic site failover or manual?

Having been involved in bringing to market many multi-site cluster solutions on most operating systems for automatic site failover, storage by itself cannot provide high availability and disaster tolerance with guaranteed data integrity for most mission critical application or database environments configured across geographically separated data centers with the goal of disaster tolerance.

The solution requires that the detection and management of a single system management domain (Oracle RAC) or single management domain (shared nothing) be done by the cluster manager built into the cluster or high availability software.

In the absence of a cluster manager managing the failure detection and taking it through the procedure of continuing (single image) or resuming (shared nothing) processing on the surviving nodes introduces risks that need to be carefully understood.

Having a metro storage cluster complementing the host based high availability and cluster software to provide non disruptive data migration, the ability to non-disruptively add and remove storage nodes, allow for storage node resource sharing to improve performance, while providing no single point of failure with data availability across the storage cluster nodes, may be an example of a storage cluster in a metro solution.

Again, interesting and great article.
Cancel

-ADS BY GOOGLE

SearchNetworking

SearchDataCenter

SearchCloudComputing

SearchCloudProvider

Close