Software Defined Data Center: Challenges
In a software defined world, you can quickly identify capacity bottlenecks and migrate tenant services.
In a software defined world, you (the IT admin) can quickly identify capacity bottlenecks and migrate tenant services for effective resource utilization. However, there are quite a few challenges on the path to realizing the true potential of Software Defined Data Centers.
On-Demand Orchestration on Existing Networks: The idea of a Software Defined Network sounds good on paper; but, how can you make an existing network flexible? You need to provision resources on-demand, expand and contract the network as demand changes, and undo changes to resources that were made during provisioning.
Multi-Vendor Environment: For technical and business reasons, you are using a diverse set of appliances from vendors like Cisco, Juniper, F5, Citrix, Brocade, VMware, Riverbed, etc., each with different versions and different APIs. Hardware independence reduces differentiation and encourages more vendors to offer the same network services, complicating the problem further.
Virtual Appliance Management: Network services that were originally limited to one or few physical appliances, will now be spread across hundreds of virtual appliances, leading to sprawl. You have to manage the virtual appliances programmatically through a policy-driven framework. For that to happen, a few difficult questions have to be answered. How do you build all the templates for virtual appliances? How do you associate the templates with a business policy? How do you manage the licenses and Images?
Most importantly, who is going to manage the virtual appliance VM? Is it the server admin, the network admin or the cloud admin?
Virtual Appliance Placement: In the physical appliance based architecture, network services are placed to honor traffic patterns such as east-west or north-south. You will have to honor many of these design considerations in the virtual appliance based infrastructure as well. Are you going to co-locate the virtual appliance VM with tenant VMs or do you want to allocate dedicated compute resources to the virtual appliance VMs?
Virtual Appliance Monitoring: You currently use many tools such as CLI, console access and traditional network management tools to troubleshoot network issues. In a multi-tenant data center that is prone to virtual appliance sprawl, you require a sophisticated solution that maps tenant services to network resources to help pinpoint the root cause of service failures.
Hybrid Architectures: Despite the advantages of virtual appliances, some network services will still be delivered using physical appliances. An orchestration solution needs to support deployments consisting of both physical and virtual network appliances.
Solution – Anuta Networks ATOM :
With the latest release, Anuta Networks ATOM delivers the industry’s first on-demand network service orchestration for leading L2-L7 virtual service appliances from multiple vendors such as the Cisco VSG, Cisco ASA 1000v, Cisco CSR 1000v, Cisco Nexus 1000v, VMware DVS, Citrix NetScaler VPX, F5 BIG-IP and Riverbed Stingray ADC. Anuta ATOM offers virtual appliance life cycle management including appliance instantiation, placement, image management, service definitions, provisioning, commissioning and decommissioning.
We look forward to seeing you at Cisco Live Melbourne from March 6th to 8th. We will demonstrate our virtual services architecture at the conference.
– Praveen Vengalam, March 4th, 2013.