White Paper Network State+ - Driving Visibility for Performance Assurance and SDN Blind SDN vs. Insightful SDN in a Mobile Backhaul Environment Extending SDN with Network State+ Introduction While the major benefits of Software Defined Networking (SDN) are relatively well understood and agreed; namely: separation of the control and forwarding planes, logically centralized software control and programmability, driving utilization, rapid innovation, manageability and infrastructure cost reduction. But the approaches to maximizing its benefits vary greatly. In some instances the approach can differ simply because of the state of the technology available to the one proposing the solution. What is not in dispute is that the more accurate the information feeding the abstracted network model, built by the SDN platform, the better the results. SDN s initial growth has largely been driven within the datacenter where network infrastructure is relatively easy to procure and provision. In this type of environment, it is largely safe to make assumptions about the expected performance and availability of the underlying network as new traffic flows are considered. This is effectively demonstrated by the fact that OpenFlow itself is very limited in its ability to deliver measurements of performance metrics such as latency. This occurs since OpenFlow is a control- plane protocol that inherently has great difficulty in accurately determining data- plane measurements. As network operators look to expand the benefits they can get from SDN beyond the datacenter and across the WAN, the need to further enhance the information available to help in decision making is increased. Decisions based on a static snapshot (reporting of data at specific times in data centers) of the expected or designed performance characteristics of the available network infrastructure cannot hope to offer the same results as one based on constantly updated measured results. For traffic flows on the network, there are currently two common solutions: 1- to use traffic counters and infer the traffic matrix for a given network as capacity planning tools have historically done, or 2- to instrument the network with a tool like NetFlow and measure the real traffic. The former solution, while adequate for long term capacity planning in the network core, does not have the accuracy that is necessary for parts of the network where traffic flows vary by the sub- second. The latter solution, until recently, has incurred too much overhead in- device processor time and management overhead to be a viable solution. Neither solutions offer a useful December 2013 Rev 1.0
answer to the question of how to report on the bandwidth availability on links that might not be operating at the designed capacity which frequently occurs in a microwave transport environment. DRIVING INSIGHT WITH NETWORK STATE+ Providing the available bandwidth in real- time is absolutely mandatory in order to effectively optimize the network. Due to the enormous amount of stranded bandwidth currently underutilized on a hop- by- hop basis bandwidth that may be stranded for redundancy purposes, because of over- engineering, or because of lack of visibility or efficient traffic engineering tools the hop- by- hop capability to see this is highly valuable. The difference between Network State and Network State+ is as follows: SDN's Network State Accedian Networks' Network State+ (TM) A must to take intelligent dynamic decisions about network op[miza[on Complete End- to- end In- mo[on picture of Transport Performance S[ll snapshots from packet counters of Top- of- Rack switches traffic distribu[on OK inside Datacenters, not sufficient across complex networks Visibility across mul[ple mediums, technologies, carriers, domains Visibility across lossy or congested paths over wide area Layer 1, 2, 3 and 4
Accedian s Network State+ (NS+) improves the situation by delivering more accurate and detailed information. Network State+ delivers a suite of parameters that provides visibility across multiple media, technology, carriers and domains. It gives added insight to account for things like latency, jitter and packet loss. Moreover, it provides visibility of available bandwidth across links and on a hop- by- hop basis. BRINGING PERFORMANCE ASSURANCE TO SDN ARCHITECTURE The current state of most mobile backhaul heterogeneous networks is essentially a network made up of leaky pipes. As illustrated in diagram #1, at the head- end you have a router blindly sending data all the way down to recipients, which may be macro- cells or small cells and then end users. These routers have very little, if any, visibility of the performance of individual links they are trying to throw traffic down to.
Diagram 1: Most of traffic is downstream; thus congestion, packet drop and delay is likely to occur in the direction from head- end to remote cell site. Traffic sent by the head- end router will be dropped further down in the network and will have consumed upstream link capacity unnecessarily. As a result, these network elements have been manually configured to a fixed type of setting like CIR or EIR. The available bandwidth may vary greatly across links because of varying atmospheric conditions, bandwidth sharing between operators, misalignment of various antennas, or interference at the spectrum level. An SDN application without awareness of the changes occurring within the network will configure the head- end router to try to send data across the links at the expected bandwidth speed, causing a lot of spillover and wasted bandwidth, especially at the top of that particular type of architecture as additional bandwidth is used for re- sent packets. What is needed in this environment is visibility. If the SDN application is aware of the current network state it can make much more informed routing decisions, throttling traffic before it hits the network bottlenecks and increasing the efficiency of the network. Diagram #2 outlines the importance of network visibility and showcases that when optimization is employed from the head- end, the optimal bandwidth can be achieved to every site, in real- time, all the time.
Diagram 2: Optimizing the traffic delivers optimal bandwidth to every site, in real- time, all the time Of course, the primary data has to come from somewhere and this is where Accedian s range of network devices and software agents are used. By offering a broad suite of hardware and software options, that can be deployed in- line or out- of- line, and that can operate with other standards compliant network elements, Accedian provides the information SDN applications require to meet their full potential. SUMMARY In a performance assured SDN based networks, applications can utilize the abstracted network model provided by the SDN platform to make informed decisions on how best to optimize their network interactions and routing requests to achieve the desired results. For that to happen, the network model must be populated with current, accurate and complete Network State information in an efficient manner thus, the role that Networks State+ has in the network. To learn more about Accedian Networks and Network State+, please visit www.accedian.com. Accedian Networks Inc. 2351 Alfred- Nobel, Suite N- 410 St- Laurent (Montreal), Quebec, Canada, H4S 2A9 Toll free: 1-866- 685-8181 2013 Accedian Networks Inc. All rights reserved. Accedian Networks, the Accedian Networks logo, antmodule, EtherNID, Fast- PAAs, High Performance Service Assurance, Performance Assurance Agent (PAA), Plug & Go, MetroNID, MetroNID GT, MetroNODE 10GE, MetroNODE LT, VeloCITY FS, Multi- SLA, NanoNID, SkyLIGHT, SLA- Meter, Traffic- Meter, VeloCITY FS, VeloCITY FS 10G, Vision EMS, VisionMETRIX and V- NID are trademarks or registered trademarks of Accedian Networks Inc. All other company and product names may be trademarks of the respective companies. Accedian Networks may, from time to time, make changes to the products or specifications contained herein without notice. Some certifications may be pending final approval, please contact Accedian Networks for current certifications.