Cisco Unified Service Delivery combines the power of the data center with the power of the network to transform service delivery and build the foundation for cloud services. Learn more by viewing the video or slides below.
Posted by Suraj Shetty at 08:53PM PST
The article “Cloud Computing Overlay for Unified Service Delivery: Delivering Infrastructure-as-a-Service” is an excellent article. Have some comments on the “Quality of Service (QoS) to meet customer service-level in IaaS offers”.
Does QoS refer to network Service QoS or Service level QoS? If it is Service level QoS for end-end, it would translate to network Service layer QoS in a particular place in the network (PIN) - Virtualized Data Center.
Cisco provides best practices QoS Design for Campus, Branch Places in network (PIN) for managing Telepresence.
Would be interested on the QoS design best practices and how they are are setup (provisioned) and shutdown (unprovisioned) as part of the IaaS offering and Cloud computing model.
Thanks, Chandramouli, for the questions. Your first was on QoS for IaaS offerings. Unified Service Delivery uses network-level QoS mechanisms as a part of ensuring overall QoS at the level of the offer or service, but it’s not the only mechanism. We also have application-fluent components that provide QoS beyond basic network-level mechanisms. The Nexus 1000V switch is one of the best examples of how network-level QoS is assured automatically and provisioned dynamically in Unified Service Delivery. This particular element of Unified Service Delivery enables a QoS policy to follow a virtual machine as it moves across the infrastructure without any intervention by the provider. This is due to VM-awareness built into the Nexus 1000V switch. We couple this with other QoS mechanisms to provide overall offer-level assurance of the quality-of-experience.
Your second comment was on dynamic provisioning throughout the network, particularly focusing on the IP NGN WAN Core, Aggregation and Access tiers. You use the example of TelePresence sessions, but I’m assuming you’re not limiting yourself to that example. In general, we will use virtualization (in the broadest sense of the term) and network intelligence to avoid reconfiguration or re-provisioning of the network elements - especially at the physical level - for specific sessions. Instead of creating separate dedicated circuits which we set up and tear down every time, we’ll employ encryption, QoS and virtual networks that are application-fluent so that the network doesn’t require reconfiguration to ensure a secure and high-quality experience to the end-user.
In this brief space, I can’t provide much more detail that that, and naturally techniques will vary depending on the nature of the service in question.
But the Unified Service Delivery solution is designed to equip service providers with a broad set of tools they can apply horizontally across their entire service portfolio to achieve the results you describe in your comments. Thanks again for the input!
Ping this URL to post a trackback:
http://blogs.cisco.com/trackback/7364/VGlUVDI5/
Previous post:
From China IPv6 Summit to Futurenet in Boston
Next post:
Architectural Innovation: A Clean Break from Tradition
Recent posts:
November 2009 Archive
Subscribing to this blog will allow you to receive entry updates to the blog via email.
To subscribe to this blog, follow the simple instructions below:
Chandramouli Venkatesan May 13, 2009
Had a question on how the IP-NGN based networks and existing Access Edge switches, Distribution switches, Core routers would work in this secure cloud computing model. ie how do we partition the virtual network for different services, setup network resources when end-user requests for it(provision physical and logical ports for example) and shutdown (unprovision/reset physical/logical ports) when the end-user relinquishes service. E.g: End-user wants to schedule the Telepresence Conference room services from home from 7pm-7am due to cost/business reasons