THE ORGANIZATIONS behind Software Defined Networks have done an incredible job focusing on the needs of carriers and of data centers. SDN is being rolled out to define the resources and configurations of carrier networks and cloud data centers, and the benefits include allowing for the rapid provisioning of services, and flexibility in setting up network configurations. That’s a great start. The next step is to dig deeper to use SDN to focus on the service needs of individual applications — and on helping coordinate resource allocation to ensure that those applications run correctly across all the component parts of the solution, which might include multiple carriers, datacenters, and cloud service providers. Creating a framework for cloud applications to communicate with the wide area network – and vice-versa – is one of the key initiatives of theCloudEthernet Forum (CEF), an industry alliance working to help scale Ethernet technology for delivering cloud services. The CEF framework defines and coordinates the interfaces between applications and networks.
For example: Consider a unified communications (UC) application, one which delivers voice calls, video conferencing, conferencing calls, and other types of messaging within an enterprise and out to its customers. UC applications are increasingly hosted by cloud providers, and require solid, reliable, predicable bandwidth on a network – not only across fast links to the customer’s enterprise campus using Internet or private links, but across the public Internet and other carriers to employees and customers working at home, in hotels, in field offices, in remote locations, in different countries and continents, on WiFi and cellular connections.
The UC application needs the ability to signal to network, DC and even other cloud service providers that the application needs specific types of connectivity, or to grow or shrink bandwidth, compute resources and storage. Similarly, network providers need to be able to tell the UC application what types of transport are available, latencies, security, and feedback information on service quality and capacity utilisation. All this communication needs to be in real time, so that if some types of connectivity aren’t available, or if there is a failure or security issue somewhere, the UC application (and end users and system administrators) can understand what’s going on and where the problem is. According to one study, fully 60% of application failures within UC applications are caused by network problems – not software issues. Since you can’t fix the problem until you find it, this type of communication is crucial to customer satisfaction and meeting SLA requirements.
Hypothetical? Not in the slightest. Competing UCC companies are collaborating today to understand the benefits of standardizing how UC applications communicate resource availability and service requirements with carrier networks and cloud data centers, not only in the CEF (which focuses on the interaction and interfaces between the different players in the end to end service) but also in other standards bodies (which look at vertical environments, like SDN controllers).
It’s time to extend standards to the cloud, and to encompass the network’s reaction to application requirements – and the application’s reaction to what is going on in the data center and the network. That’s where the OpenCloud Project comes in.
Launched by the CEF in March 2014, the project is focused on application performance management, cloud security and traffic load balancing in a real-life test and standards-development environment, the first such in standards history. Since its announcement, the CEF has been working iteratively on the first iteration of a reference architecture framework in the OpenCloud environment, and the first batch of use cases. For example, UC is one of the use cases the CEF are currently developingto develop on the OpenCloud test bed.
The standards framework was approved by the CEF’s board of directors in late January 2015, and an executive summary should be available in March. In April at Interop in Las Vegas, the CEF plans to demonstrate how the framework can be used for end users to request application resources from cloud service providers over a network… and also for cloud service providers to request resources from networks. The solution must work both ways to be truly effective.
The unified communications use case is only the start. The standardized framework and use cases encompass other applications, such as security, reporting and service management, provisioning, SaaS productivity tools like Microsoft’s Office 365, Salesforce.com, Amazon’s newly announced WorkMailand so on. The sky’s the limit. Or should we prefer to say, with the right technology and standards, there are no limits to what applications, networks and data centers can do in the cloud.