‘Anything-as-a-service’ is the hot button these days. Certainly that term is associated with ubiquity and simplicity for most people and that is also what we have in mind here at Enterasys as well. Simplicity drives operational efficiency up, cost down and is the basis for higher agility. At Enterasys we grew up with switching/routing simplicity in mind, so we know a thing or two about these topics. We recently released new products and capabilities in data denter with Fabric Routing and IP Host Mobility – simplified and optimized routing for east/west and north/south traffic, respectively. In our upcoming releases we will be taking these capabilities one step further towards an even more simpler solution – yes, that is actually still possible! That degree of simplicity drives one of the key values of our OneFabric Data Center architecture even higher.
So what does this mean? It means the need for simplifying the distribution of routing (service) everywhere in a larger network is crucial. Let me explain why. Fabric Routing is already a mechanism that provides distributed routing in SPB (Shortest Path Bridging) and VSB (Virtual Switch Bonding) based switch/router designs. This is to address the need for maximum throughput, lowest latency and optimized traffic flows inside the data center fabric for both switched and routed traffic. However, it does require some configuration on the participating fabric switch/routers. In a typical enterprise data center this is a relatively easy task because of the low number of switch/routers, but if you intend to expand this type of design into very large data centers (or even the campus network) then another level of simplification with “zero-configuration” is required.
But why would you want to deploy SPB in a large campus network? The IEEE SPB standard makes it possible to design large Layer 2 networks that are stable and scalable. In the data center this is an obvious design option to enable VM mobility. For certain applications in the campus network SPB also becomes a requirement. At Enterasys we are continuously asked how to provide selected Layer 2 services across the campus LAN using our Virtual Private Port and Ethernet Services – this is a requirement for many of our customers! PCI regulations and other considerations drive network architects toward solutions that address this requirement. As well, if one takes the IP mobility challenge of Wireless LAN (WLAN) that arises as you move away from “controller tunneled” network designs (as a huge amount of traffic from high density 802.11n and 802.11ac WLAN networks will force you to do so) to a “bridged at the access point” architecture, then suddenly those designs become a reality. We are also seeing that IP addresses are no longer an indicator of the location for a device or user. Users and their devices are scattered across the network in multiple different subnets and VLANs that are distributed across various switches and access points. So where do you route the traffic between those networks? What is the best placement for a router in such a design?
The answer is: EVERYWHERE! Each switch becomes a router in the “domain.” Today, with an Enterasys network, whether at the data center or campus, the best path is always selected because it is done with Fabric Routing. As I pointed out earlier in this blog, today using Fabric routing requires some level of configuration on the participating switch/routers. For a campus where you route at the distribution or core that is perfectly acceptable. But when we have 100’s or 1000’s of switches in a campus network that all should route to optimize the traffic pattern in a larger Layer 2 design, these switches all need to be auto-provisioned with “zero-configuration”. This means pushing Routing from the Routers to a Service everywhere!
Think this sounds exciting, interesting and different? Sounds RaaSnable to me! Stay tuned for a Technology Strategy Brief in the coming month.
The post Routing as a Service – does this become automatically “RaaS”? appeared first on Extreme Networks.