Unified Fabric and the Network Value of the Data Center
Deliver information fast…really fast!
Never go down, ever!!
Enable great user-experiences, regardless of where the user is on the planet.
Those are the table stakes requirements of today’s Data Center networks. Do that and business leaders will ignore it like a referee that never makes a bad call in a sporting event. Click-to-“Right Now”, Borderless, 24x7x365 networks. No problem. We can do that.
But with the pace of technology change in today’s Data Center, so much more is being expected of the network:
- Unify multiple types of traffic onto a single infrastructure, preserving the delivery characteristics each had on isolated networks (eg. Storage, Video, Market Data, etc.)
- Consolidate multiple silos of Compute and Storage into large, shared pools of resources that can be dynamically allocated as needed.
- Create logical segmentation and isolation of resources on top of those shared pools of resources.
- Enable new traffic patterns (server-to-server, live migration of VMs) without reducing the control available in previous architectures
- Reduce or eliminate the concept of geography as it relates to Data Center resources.
- Create a conduit to allow services to be delivered through the best means possible, internal or external.
So how we deliver all of these next-generation capabilities across the network within the Data Center Business Advantage architecture?
Consolidate the Traffic: Cisco has done this many times before. First it was Mainframe (SNA) and TCP/IP traffic traffic on the same network. Then it was Telephony (VoIP) and Video on the same LAN/WAN. And now we’re seeing the move of Storage traffic onto a Unified Fabric. The first step is a migration to 10Gb Ethernet for higher performance, with the next step being an introduction of Fiber Channel over Ethernet (FCoE) to consolidate storage infrastructure. By consolidating traffic onto a unified fabric, customers can reduce physical infrastructure (cables, cable racks, patch panels), reduce power and cooling (few devices, better airflow) and create the foundation to enable pools of resources within their Data Center.
Enable Dynamic Application Awareness: Just consolidating traffic onto a 10Gb Ethernet network is not enough. That network must be able to dynamically adapt to the needs of multiple applications and types of traffic. A dynamic network uses Quality of Service (QoS) to dynamically allocate bandwidth and reduce risk to all application. By embedding the application awareness into the network, new applications can be easily deployed and centralized policy can be deployed to ensure performance for all applications.
Virtualize and Segment the Network, End to End: As the Data Center evolves, most IT architects agree that it needs to adopt many of the characteristics of the largest public cloud computing environments. Key to this is the ability to provide services for multiple tenants, whether those “tenants” are customers (for a Service Provider), or divisions/groups (for an Enterprise). Cisco’s Virtualized Multi-Tenant Data Center (VMDC) architecture provides a reference for customers looking to extend virtualization and segmentation services from the core network through to the virtual machine (VM), inclusive of storage. This architecture is an excellent compliment to integrated computing solutions such as Vblock and Secure Multi-Tenancy (SMT).
Overlay Next-Generation Layer 2 Technologies: Enterprise Data Centers went through a massive transition in the mid-to-late 1990s as “Layer 2 switching” technologies came into the market, offering greater throughput and performance to applications. But Layer 2 also brought with it a lack of control and added complexity for IT architects. As a result, a three-tiered architecture (mixing Layer 2 and Layer 3 technologies) became dominant within the Enterprise. Fast forward to today and virtual machine (VM) mobility is causing IT architects to rethink those L2/L3 assumptions as mobility demands bump into network boundaries. Cisco addresses these challenges with two innovative technologies, FabricPath and Overlay Transport Virtualization (OTV). FabricPath allows infrastructure architects to add broader Layer 2 flexibility within the Data Center to enable greater server-to-server performance and VM mobility. OTV simplifies the ability to extend Data Center boundaries across locations, allowing new ways of thinking about challenges like Business Continuance. So now infrastructure architects can deliver the simplicity and efficiency of advanced Layer 2 networks without abandoning the Layer 3 control and operations they have enjoyed for many years.
Rethink the Boundaries of the Data Center: The combination of enhanced Layer 2 technologies from Cisco, live VM migration, and storage enhancements from eco-system partners EMC and NetApp allow customers to begin thinking about Application Mobility in ways that were never possible before. The Data Center is no longer just a single location, but rather a cluster of resources that are not limited by location. Cisco IT is already beginning to enable this functionality in our own Data Centers. As network, virtualization and storage technologies evolve, the boundaries of the virtualized Data Center will expand to tens of thousands of kilometers (asynchronous distances). And once you can wrap your head around the concept of Data Centers extending beyond the existing walls, it’s not that far of a leap to start thinking about leveraging public and private (internal) services in conjunction. Yep, a Hybrid Cloud model that allows the Enterprise to stay in control.
Demands on the network, especially within the Data Center, have never been higher. But this is not a unique phenomenon, it’s happened every time a new transition of previously silo’d traffic moves onto the IP network.
OK, so enough with the technical details…
….so why should you care about Unified Fabric as a Business Decision Maker?
- Because today’s businesses can no longer afford the real-estate, environmental or operational costs of running isolated, business-unit centric networks.
- Because application performance, risk management, compliance and security no longer have to be mutually exclusive on your Data Center network.
- Because computing models will begin to mimic financial models and “follow the sun” (or “follow the moon” for lower energy costs), so failing to establish a virtual data center strategy will leave companies at a competitive disadvantage.
- Because a unified network fabric, paired with a unified computing model will accelerate the ability for IT to deliver applications to business users and customers.
Cisco’s Data Center Business Advantage architecture provides a framework to understand how today and tomorrow’s network challenges can be solved. Network challenges for application performance, application mobility and application risk management.
Now that we’ve looked at the location and deliver of your critical business applications, let’s look at how to add flexibility on how to improve them.Tags: