Skip to content 
This is the second of five posts on The 5 Tuple Challenge.
With the growth of cloud computing, customers have different types of requests. Different use-cases, such as hosting a website, scalability tests or disaster recovery plans, could result in different requirements for the network topology and network services. In a multi-tenant environment, all customers must share the same cloud infrastructure.
The cloud administrators have two questions on their mind: Are the tenants satisfied? How can the admins generate more revenue while using the infrastructure efficiently?
You can’t improve what you don’t measure.
So, the cloud administrators need real-time visibility of their infrastructure. They need to know which network service has been ordered by which specific tenant and where it has been provisioned in the infrastructure. Also, they need to know what the most commonly requested network services are along with the lifetime of those services.
To ensure customer satisfaction, they need availability information. For example, how long has a data center been up and running? Were there any link failures or alerts? How long does it take to provision a network for a tenant?
In addition, they need to monitor the overall resource utilization. For example, they need to know how many VLANs, VRFs, firewall contexts and load-balancer contexts are used at the pod, data center and region levels. They need historical usage information to justify investment in infrastructure as well as to forecast customer demand and procure hardware accordingly.
On top of that, cloud administrators need to optimize the usage of the existing infrastructure.
Since services need to be delivered on-demand, they need a solution to provision those resources intelligently.
Let me explain with an analogy. Consider the typical convention center that has a huge auditorium. Based on demand, walls are setup to prepare many small rooms or a few big rooms. The walls can be setup or torn down instantly.
Since services need to be delivered on-demand, they need a solution to provision those resources intelligently.
Similarly, cloud administrators need flexibility to offer a different mix of network services – for example, Gold, Silver or Bronze – based on pricing and demand.
Additionally, cloud administrators may want to set limits on the number and types of services. For example, they may want each data center to support at least 500 gold services or they may want to limit bronze services to below 10000 tenants. Implementing such business intelligence in an on-demand environment is very complex and costly.
Anuta Networks nCloudX offers comprehensive real-time network service monitoring for cloud deployments. nCloudX detects problems related to performance and availability that can impact both physical and virtual service elements. Detected problems are mapped to the service instances and tenants. This helps cloud administrators communicate effectively with tenants and isolate the root causes of problems.
nCloudX also discovers existing capacity in the data center to provide an accurate view of the number of the services of a given type that can be provisioned. Administrators can define resource pools, set logical limits on each data center and write policies such that a particular data center supports a minimum number of services.
nCloudX will optimize resource provisioning according to policies while maximizing infrastructure utilization.
They can even decide the range of VLANs, VRFs, firewall contexts and Load-balancer contexts available for dynamic provisioning. nCloudX will optimize resource provisioning according to those policies while maximizing infrastructure utilization.
Finally, nCloudX delivers detailed reports about data center resource usage and performance, tenant service usage history and overall network service demand.
This will enable cloud providers to plan their data center expansion while ensuring customer satisfaction.
In the next post, we will discuss challenges related to on-demand network service provisioning.
I look forward to seeing you at VMware Partner Exchange in Las Vegas from Feb 25 to 28.
– Srini Beereddy, February 19th, 2013.
