Preview only show first 10 pages with watermark. For full document please download

Data Center Applications And Mrv Solutions

   EMBED


Share

Transcript

Application Note Datasheet Data Center Applications and MRV Solutions Introduction For more than 25 years MRV has supplied the optical transport needs of customers around the globe. Our solutions are powering access networks for customers around the globe. Our unique expertise in data center applications and flexible optical transport solutions give our customers the options to create networks that match their specific requirements. The 8 applications described below provide an overview of ways that the OptiDriver, MRV’s optical transport platform, overcomes optical transport challenges for our data center customers. Peering PoP Offload Problem: Because many data centers are located in urban centers, the costs for real estate, rack space, electricity, cooling, cross-connect, labor and other charges can be high. When data center resources are needed these costs can be prohibitive for some users and applications. Solution: Locate the data center in more distant locations where these cost are less expensive and connect to the urban data center using dark fiber and DWDM. Benefit: This provides more space for servers, storage and applications with much lower operational costs. Energy for cooling can be minimized through site locations that use natural cooling and hydroelectric power generation can provide a clean energy source. Description: The OptiDriver is designed from the ground up to reduce capital and operational expenses. It uses a flexible shelf and component architecture that dramatically lowers initial and ongoing hardware costs, simplifies operations, and minimizes power and space requirements. With a single system and minimal number of components, network designers can serve the widest variety of applications. Designed for access and metro DWDM and CWDM connectivity, OD supports any optical transport requirement with unsurpassed functionality including Ethernet, OTN, TDM, Fibre Channel, ESCON/FICON, Infiniband, and other applications. The system is designed for ease of deployment, rapid circuit activation and certification, acceleration of revenue generation, and return on investment. Software-based management features minimize both initial and recurring costs. 1 Application Note Datasheet 100G Access Problem: One customer needs 100G access to the data center or many customers from similar locations need 10G access to the data center. Fiber is available but there may be limitations on the lambdas available on the DWDM network. Solution: Use a 100G CPE that can support one customer at 100G or aggregate many customers at 10G onto any single 100G wavelength of the ITU grid. Benefit: Customers get the data center access they need and the data center gets new revenues while preserving optical spectrum using only one wavelength. Description: MRV’s OptiPacket OPX-1 metro service edge platform, has a fully non-blocking switch fabric. It has the capability of supporting up to 20 ports of 10G each and 2 ports of 100G each. Or with a separate line card it can support 4 ports of 100G. This provides significant flexibility when used with individual customers or in multi-tenant buildings. Supporting twenty customers each receiving 10G of guaranteed bandwidth is easy for the OPX-1. Each customer is assigned to one of the two 100G ports for full non-blocking throughput. However, some customers want to create a redundant link with the second 100G port. These customers typically require a diversely routed fiber to the second port. This way they have a full 100G connection in the event of a fiber cut on their primary 100G port. The full 800G switching capacity is available to the users. Full line rate traffic can be provided to and from each port so guaranteed services are delivered according to the SLA. However the OPX-1 can also be configured to deliver best effort services as well. When the port is fully committed to guaranteed services, then best effort traffic will be dropped at ingress or at egress. When high priority traffic is absent then bandwidth is available for best effort traffic. Another use of the OPX-1 is for 100G access to a single customer. The bandwidth requirements for large campuses, government, enterprise and research universities are increasing rapidly. The OPX-1 is purpose built to give fine grained control and manageability to services running at 100G. At 100G, the OPX-1 supports CFP optics with standard link distances of 40km. For 10G interfaces the OPX-1 supports SFP SFP+, Tunable SFP at 1G, 10G and 40G. All MRV interfaces also provide digital diagnostics so users get performance data from the optical module. Low Latency Interconnect Problem: Some applications require extremely fast response time (low latency) to perform correctly. In order to get the required performance from the network, latency must be reduced from each network element and process. Solution: Fast response time can be achieved by using these design criteria: 1) Use an optical transport solution that has low latency components and low overall systems latency. 2) Design the network for low latency by following low latency design rules. 3) Finally, keep the fiber distances as short as possible. Benefit: These three strategies will minimize latency in the optical transport and in conjunction with low latency servers, storage and applications, will deliver the fastest response available. 2 Application Note Datasheet Description: Finance in general and high frequency trading in particular requires the lowest latency possible. There are also applications in high speed computing and some client server application that also require it. The OptiDriver solution has the overall lowest latency in the industry. In addition, some components, including the 100G transponder has orders of magnitude lower latency than competing offerings. We measure latency in nanoseconds and the competition measures latency in milliseconds and the difference can be very significant for customers concerned with supporting low latency applications. Disaster Recovery Problem: Critical information stored at a data center must be replicated and stored at a separate location to ensure business continuity or to meet regulatory requirements. Solution: Create an optical transport link to a 2nd data center at a remote location to replicate the mission critical data. Benefit: Ensures that an enterprise can function even in the event of a catastrophic failure of one data center and meets regulatory requirements for financial and insurance purposes. Description: The OptiDriver transport network can be built in a ring topology. Instantaneous optical switching in the OptiDriver can reroute traffic from east direction on the ring to the west direction in a few milliseconds. Users on the network do not even know they have been redirected to a back up data center. Any users attached to the OptiDriver network at any point on the ring are automatically redirected. This layer 1 switching can be invoked independently of any Layer 2 or Layer 3 mechanisms. Mesh Interconnect Problem: Rapid growth with addition of buildings and customers may prevent data center operators from knowing the exact locations they will need to connect. When these locations are defined then the optical transport network needs to be flexible enough to support connections between and among all the metro data center locations both new and old. Solution: Build an optical transport network capable of adapting to the changing environment. This requires flexibility of data rate, protocol, wavelengths, management and other parameters. Benefit: Pay for only those links that are needed and gain the peace of mind and flexibility to create any other links to local or distant locations as the needs arise without limitation. Description: The OptiDriver platform is designed for maximum flexibility, low cost, low power and small footprint. It also offers data center operators the options to add modules to connect buildings near or far with almost any protocol and in any topology. These features mean that operators need only buy what they need now, for their current needs, and then add modules for the requirements that arise in the future. No matter what metro optical transport application you require, the OptiDriver offers the most functionality and best value on the market. It supports more protocols and data rates than competing solutions. Resilience is built in with multiple redundancy schemes and the best granularity to ensure reliability of services. Advanced features like ROADM and OTN provide the flexibility to build any network topology and service. 3 Application Note Datasheet Management capabilities enabled by an intelligent control plane based on SDN and GMPLS technology and supported by ProVision offer efficiencies that deliver simple and reliable network scalability. All these features combined with the smallest service to footprint ratio and most efficient power consumption in the industry make OptiDriver the platform of choice for service providers looking to build competitive, high-performance, next-generation optical networks. Multi Protocol Interconnect Problem: Cross-connection between multiple communication service providers and customers at a data center meet-me room or Internet Exchange often requires support for a number of common protocols including: Carrier Ethernet 2.0, SDH, SONET, Fiber Channel or Infiniband. Solution: Use an optical transport system that has a wide arrange of options including, size, protocol, number of ports, impairments compensation, switching and optics to enable connectivity for many protocols and data rates with modules that enable scale from a few ports to many. Benefit: This strategy ensures that no matter what customers may come to the data center in the future, there is a simple and inexpensive method of connecting them to any service provider. Description: OptiDriver supports a wide range of signal speeds and formats – from T1/E1 to 200G – with a minimum number of physical cards through the use of innovative multi-rate transponders. By reducing the number of parts overall, OD reduces sparing costs and operational complexity while increasing network flexibility. Pushing the definition of a software-defined network to the optical layer, OptiDriver’s flexible transponders allow service providers to make changes at the optical layer without requiring changes to the physical infrastructure. These flexible transponders can also be configured to support multiple service types, including dual-transponder, client/line protect, and multicast without requiring additional hardware. The design of the transponders combined with technologies such as low-latency dispersion compensation modules ensures ultra-low latency throughout an OptiDriver network. MRV’s 100Gbps solutions have been shown to have two to twenty times lower latency than competing systems. Demarcation for Cloud Problem: Wavelength services, dark fiber and FaaS (Fiber as a Service) have no method to check the health or performance of the fiber link or the services running on that link. Solution: Providing an optical demarcation device at the customer site provides the opportunity to collect performance and fault management data. Benefit: Service providers can still sell their current services but they get the added benefit of reduced troubleshooting costs and the ability to guarantee and prove overall SLA compliance. Description: OptiDriver modules can be configured to provide information on optical performance including Digital Diagnostics that include: Tx/RX, OSNR, power, voltage, temperature and others. In addition, for modules supporting FEC additional performance variables such as BER, frame alignment and others are available. Carrier Ethernet service performance parameters can be collected, analyzed and reported, even if the service provider is not providing a Carrier Ethernet service. If Ethernet packets are traversing the link then these performance characteristics can also be made available. 4 Application Note Datasheet Parallel Ultra High Bandwidth Problem: Data centers may require massive bandwidth on static links between metro DCI locations. Solution: MRV provides the densest 100G/200G transport at the lowest cost with few features. Benefit: High bandwidth point-to-point links make the data centers appear to be one virtual data center so that Virtual Machines, processing, data and storage can be moved at will among the different physical locations. Description: OptiDriver supports up to 8 x 200G links in one chassis with a total capacity of 3.2T. These links can be installed and tested all at once or added over time as requirements demand. OptiDriver modules exist for muxponding multiple 10G, 40G and 100G ports over the 200G links. This makes connection of switches and routers simple and efficient. Conclusion MRV solutions for optical transport in the data center market offer the flexibility to meet any application. As the previous examples demonstrate, MRV has the low latency, small footprint, low power and low cost needed for today’s operators. Whether it is managed by MRV’s optical service platform, ProVision, or with another network service orchestration platform, the OptiDriver platform is the industry’s gold standard for data center interconnect. MRV operates worldwide sales and service offices across four continents. Contact us at [email protected] www.mrv.com All statements, technical information and recommendations related to the products herein are based upon information believed to be reliable or accurate. However, the accuracy or completeness thereof is not guaranteed, and no responsibility is assumed for any inaccuracies. Please contact MRV Communications for more information. MRV Communications and the MRV Communications logo are trademarks of MRV Communications, Inc. Other trademarks are the property of their respective holders. MRV-AN-Data Center_06-19-16 Copyright ©2016 MRV Communications, Inc. All Rights Reserved.