Offering greater agility, and better security than traditional client-server architectures or public clouds, it’s little wonder that a growing number of organisations are investing in virtualised, automated data centres and private clouds.
While server and storage elements in the private cloud are mostly automated, the network is likely still provisioned and configured manually. So taking full advantage of the promise of private cloud requires scalable network control to make sure deployments aren’t held back by legacy processes.
As an IT department’s private cloud infrastructure matures, it will go through a series of phases, the first of which involves pilot projects. It’s at this stage that IT teams will use non-critical applications and workloads to test out their cloud’s design and infrastructure.
With the confidence and experience gained from having done so, they will then move on to a second “production” phase during which they will move one or a few business-critical workloads on to the private clouds.
The final scale-out phase sees the transition to geographically-dispersed private cloud environments in multiple datacentres, possibly including multi-vendor cloud platforms.
Every step in the process of deploying a private cloud must be in sync or deployment can prove risky for a business, regardless of the size and scope of the project.
One major challenge that must be addressed is the disparate groups dealing with private clouds. Typically, the server team handles the virtualisation component, but another team deals with the network aspects. So now, for example, then network team may have a lack of visibility into virtual machine (VM) resources as they’re created and destroyed, making it difficult to track and manage the massive spike in new instances.
Without visibility, there’s little point in trying to comply with security and audit policies because there is no accurate information on which IP addresses and DNS records are assigned to which VMs at any given time. Factors such as locations, applications, and users need to be tracked for VMs as well as networks, IP addresses, and DNS zones. While most server admins will have access to part of this information, networking teams will not. Often still using manual methods to react to the creation and deletion of VMs, their responses will tend to be slow.
Slow and costly
Unreliable DNS, DHCP and IP address management services – collectively known as DDI – can threaten an organisation with potentially costly network outages.
What’s more, private clouds are only as fast as their slowest component. When building a private cloud, it’s important to consider core network services such as assigning IP addresses and DNS records so that VMs can easily be commissioned and decommissioned in a matter of moments.
Private cloud’s promise of rapid delivery can be inhibited by the hours, or even days, that it can take to manually provision DNS records, and IP addresses in a virtual environment. The process of manual IP reclamation can be inaccurate and inefficient, and can result in a sprawl of unused IP addresses and DNS records. In addition, if IP addresses of VMs are used for the purpose of billing internal “customers,” then it’s possible the use of manual processes will lead to inaccurate charges. And potential IP address conflicts resulting from just a few small keystroke errors could cause significant downtime in the private cloud environment.
Private clouds running critical workloads, or spanning numerous geographical locations, require highly available DDI services to provide scalability and resilience. Limits on the scalability of an organisation’s network may prevent the deployment of additional tenants and VMs required to meet the demands of its growth.
Automation, visibility and integration
To stand a greater chance of success in its private cloud deployment, an organisation must understand and give urgent consideration to critical factors such as those mentioned above. An approach based on principles such as automation, visibility, and integration will allow an organisation to take more effective control of its private cloud deployment.
In most successful private cloud deployments today, the management of storage and compute is heavily automated, supporting the agile delivery of low-cost services to lines of business, delivering tangible benefits across the organisation.
About the Author
Arya Barirani (@abarirani) is VP of Product Marketing at Infoblox. An 18-year marketing veteran, he has a record of success in driving growth and capturing market share. He has held leadership positions at Symantec, HP, Mercury Interactive, Veritas Software, and Computer Associates International.