RSS FeedWhite Papers

White Paper Download

Dedicated networks for high-performance, predictable, and resilient IP storage

Dedicated networks for high-performance, predictable, and resilient IP storage

Dedicated networks for IP storage

Category: Big Data

Date: , 16:12

Company: Brocade

People often think of the IP/Ethernet network as one large, integrated network for all end hosts, appliances, servers, and IP storage. This is likely because network developers have managed to converge, over the past two decades, various disparate networks into a common IP/Ethernet network.

Download this white paper to find out how you can ensure autonomous administrative control and operation.

AT A GLANCE DATA CENTER HIGHLIGHTS • Dedicated networks for IP storage provide predictable performance, ensure proper levels of security, contain failure domains, and maximize uptime—all critical attributes for today's mission- critical IP storage applications. • Brocade VCS Fabric technology and Brocade VDX switches provide a robust network infrastructure ideally suited for dedicated IP storage environments. • Hallmarks of the Brocade VDX switch family include unsurpassed automation, a load-balanced, multipath architecture for maximum link efficiency and resiliency, and deep buffers to handle bursty storage traffic. Dedicated Networks for High-Performance, Predictable, and Resilient IP Storage People often think of the IP/Ethernet network as one large, integrated network for all end hosts, appliances, servers, and IP storage. This is likely because network developers have managed to converge, over the past two decades, various disparate networks into a common IP/Ethernet network. This process began with voice convergence and continued with similar work in video and storage. The business drivers behind the push toward a converged IP/Ethernet infrastructure are straightforward. After all, why would an organization not want a single IP/Ethernet network interconnecting everything? The economies of scale should reduce capital costs and having fewer networks to manage should minimize complexity and reduce operational costs. And if organizations need to segment the IP network, they can do so with VLANs, VRFs, or overlays on the single shared IP/Ethernet network, right? While the theoretical benefits sound compelling, in practice organizations seldom take this approach for numerous business and technical reasons. This also is the case with IP storage. DEDICATED IP STORAGE NETWORKING EXAMPLES The modern medium to large-scale data center typically has many separate IP storage networks, including: • Backup network: For example, an IP-based tape/virtual tape/deduplication network driven by the need to minimize RPO/RTO thresholds. • IP storage back-end network: To isolate node-to-node communications within a storage cluster. • iSCSI block storage network: Segregated primarily for storage performance reasons. According to Gartner, 30 to 50 percent of iSCSI deployments use a separate dedicated network for performance. • vMotion network: VMware's best-practices guide recommends a separate network for vMotion (including Storage vMotion). • Object store: For data centers in which capacity is measured in petabytes or larger units, content is unstructured or semi-structured, and the need is for scale-out with eventual consistency. The object store is optimized for cost and BROCADE scale, as opposed to performance and transactional consistency, and is often deployed and managed separately as part of an analytics project. • Virtual infrastructure storage: To provide dedicated IP storage for Virtual Machines (VMs) and their associated data—a common scale-out NAS use case. • Replication network: For example, distributed storage technologies in which replication is critical for redundancy and failure handling. A dedicated network can be deployed as either a physically separate network or as a separately managed network. The need for dedication varies and can include performance and service-level guarantees, contained failure domains, security, or isolated change control and span of control. For example, if virtual infrastructure storage is deployed to provide VM images and their associated data for a server farm, the best practice is to have a physically separate network (see Figure 1) for management, performance, security, and failure domain reasons. Using this model, end hosts or other appliances do not need to have network access, and Internet connectivity is not needed. A physically separate network also provides a separate change control domain and allows for Service Level Agreements (SLAs) specific to the virtual infrastructure storage use case. Conversely, a new analytics pod introduced into an existing shared network can be deployed as a separate pod, with different compute, storage, and networking hardware. The pod also can be deployed and managed by a separate team, and requires connectivity only into the spine of the existing data center network. In this scenario, the deployment decision is often driven by the Line of Business that owns the application or workload, not just the IT infrastructure owner. DEPLOYMENT STRATEGIES AND BEST PRACTICES The following examples illustrate ways to deploy dedicated IP storage networks using industry best practices. Virtual Infrastructure Storage Consider a server farm that needs access to a large library of VMs. In this case, all virtual servers in a data center might rely on NFS for their boot drives, application drives, data, and more. As such, a network outage could be catastrophic to all business logic/applications simultaneously. This is in stark contrast to traditional, classic NAS deployments as a simple file share in which outages are far less impactful. The most recent best-practices document from VMware states: Private Network vSphere implementation of NFS supports NFS version 3 in TCP Storage traffic is transmitted in an unencrypted format across the LAN. Therefore, it is considered best practice to use NFS storage on trusted networks only and to isolate the traffic on separate physical switches or to leverage a private VLAN. All NAS-array vendors agree that it is good practice to isolate NFS traffic for security reasons.1 In this deployment, the NFS environment is analogous to a Storage Area Network (SAN). It should be modified only when absolutely necessary, and should have a separate change-control decision-making process. Traditionally, networks experience problems for many reasons, including misconfiguration, software defects, and human error. Physically isolating the environment ensures that these disruptions are minimized to the highest degree possible, thus ensuring maximum application uptime. This protection cannot be provided simply by isolating this traffic to specific VLANs on a shared infrastructure. Separate physical switches are the best practice. In summary, this use case is best served by a separate physical network for security, span of control, and uptime. Figure 1. A physically separate IP storage network delivers predictable performance, low latency, and high availability. 1. VMware, Inc., Best Practices for Running VMware vSphere® on Network-Attached Storage (NAS), Technical Marketing Documentation V 2 .0, January 2013. iSCSI Block Storage Networking The next example is the iSCSI SAN. Unlike NAS, these deployments often use separate dedicated network switches as noted in the Dell EqualLogic Configuration Guide: Note: It is recommended to use a physically separated network dedicated to iSCSI traffic that is not shared with other traffic. If sharing the same physical networking infrastructure is required, then use Data Center Bridging (DCB) for EqualLogic SAN.2 The same guide also states: Several switch vendors may provide additional link aggregation options that are completely proprietary or may be extensions to one of the two previously mentioned solutions. In most cases, this type of link aggregation solution is designed to reduce or eliminate the need—and the overhead—of the Spanning Tree Protocol that must be used in the two previous options. If available, these proprietary options should be considered. They may be very useful in allowing the network administrator to create a more efficient multi-switch Layer 2 network infrastructure for a SAN.3 Backup Network Dedicated backup networks are generally considered a best practice. Although the value proposition of 10 Gigabit Ethernet (GbE)—network consolidation—includes collapsing the general-purpose LAN with the backup network in the data center, many organizations have found that this adversely impacts recovery point and recovery time objectives. Thus, they maintain separate backup networks to ensure optimal RPO/RTO. The EMC Data Domain/NetBackup best-practices guide states: By segregating NetBackup media server and storage unit traffic from other network traffic, potential contention issues are limited to backup and recovery jobs. Known available bandwidth can be managed to achieve aggressive data protection and recovery service levels. A scalable infrastructure has been established in case data protection network bandwidth requirements change over time. While not always possible based on customer requirements and pre- existing NetBackup media server and network infrastructure deployments, the use of a dedicated backup network is preferred when compared to mixed-use network configurations.4 Brocade VDX 8770-8 Figure 2. Brocade VDX switch portfolio BROCADE VCS FABRICS FOR DEDICATED IP STORAGE NETWORKING Brocade® VCS® Fabric technology eliminates Spanning Tree Protocol (STP) to deliver active-active links, doubling network efficiency and improving resilience. This flat, multipath, deterministic mesh network is ideal for IP storage environments. To meet the challenges of dedicated IP storage networks, Brocade VDX® switches powered by Brocade VCS Fabric technology provide the following benefits: • A highly automated and simple-to-deploy solution: VCS Fabric technology and Brocade VDX data center switches are self-provisioning and self-healing, delivering a 50 percent reduction in operational costs. • Predictable performance: Non-blocking, multipathing at network Layers 1-3 provides the industry's best and most predictable network utilization. • Deep buffers: Brocade VDX switches offer the industry's deepest buffers to handle bursty storage traffic and minimize latency and packet drops. • A solution purpose-built for next- generation data centers: Chassis-based HA, ISSU, and fixed-configuration redundant power supplies and fans provide high availability. The portfolio of Brocade VDX switches provides Ethernet storage connectivity for FCoE, iSCSI, and NAS storage solutions within a single product family (see Figure 2). IT organizations can protect their Fibre Channel investment by connecting Fibre Channel SANs to Ethernet fabrics with the Brocade VDX 6730 Switch. For additional information about VCS Fabric technology, read the white paper An Introduction to Brocade VCS Fabric Technology. Brocade VDX 6710 Brocade VDX 6730-60 Brocade VDX 6740 2. 3. 4. Dell, Dell EqualLogic Configuration Guide, Version 14.3, October 2013. Ibid. EMC Corporation, EMC Data Domain Boost for Symantec NetBackup Open Storage Best Practices Planning, 2011. AT A GLANCE SUMMARY Dedicated, private networks for IP storage have become a commonly accepted best practice in many enterprises. While the applications and business reasons may vary, a common thread is the need to support the application or workload in the most effective and reliable manner possible. The goal is to ensure autonomous administrative control and operation, as well as a tight coupling of the application, compute, storage, and network in order to achieve the objective— whether it be improved management, performance, contained failure domain, security, or span of control. ABOUT BROCADE Brocade networking solutions help organizations achieve their critical business initiatives as they transition to a world where applications and information reside anywhere. Today, Brocade is extending its proven data center expertise across the entire network with open, virtual, and efficient solutions built for consolidation, virtualization, and cloud computing. Learn more Corporate Headquarters San Jose, CA USA T: +1-408-333-8000 European Headquarters Geneva, Switzerland T: +41-22-799-56-40 Asia Pacific Headquarters Singapore T: +65-6538-4700 © 2014 Brocade Communications Systems, Inc. All Rights Reserved. 04/14 GA-AG-497-00 ADX, AnylO, Brocade, Brocade Assurance, the B-wing symbol, DCX, Fabric OS, HyperEdge, ICX, MLX, MyBrocade, OpenScript, VCS, VDX, and Vyatta are registered trademarks, and The Effortless Network and The On-Demand Data Center are trademarks of Brocade Communications Systems, Inc., in the United States and/or in other countries. Other brands, products, or service names mentioned may be trademarks of others. Notice: This document is for informational purposes only and does not set forth any warranty, expressed or implied, concerning any equipment, equipment feature, or service offered or to be offered by Brocade. Brocade reserves the right to make changes to this document at any time, without notice, and assumes no responsibility for its use. This nformational document describes features that may not be currently available. Contact a Brocade sales office for nformation on feature and product availability. Export of technical data contained in this document may require an export license from the United States government. BROCADE

You must have an account to access this white paper. Please register below. If you already have an account, please login.


Not registered?Register now

Forgot password?

White paper download

We use cookies to provide you with a better experience. If you continue to use this site, we'll assume you're happy with this. Alternatively, click here to find out how to manage these cookies

hide cookie message

ComputerworldUK Knowledge Vault

* *