Originally Developed To Provide Greater Performance In Direct-Attach Environments
Note to Presenter: View in Slide Show mode for animation. This slide depicts the evolution of storage networking technologies. This presentation will cover the benefits of SAN versus direct-attached storage, some of the applications like backup and replication that are made easier in a SAN environment, and then we’ll talk about the Fibre Channel over Ethernet protocol.
Let’s review a timeline of adoption. As you can see, it took Ethernet 20 years to really become widespread, whereas 10 GbE went from standard to widespread adoption in only a matter of 7 years. Events that will help drive widespread adoption of FCoE are things like FCoE embedded in storage arrays, FCoE embedded in SANs and LANs, FCoE embedded in servers, and finally, support for Unix. 2011 has brought 16Gb FC, and 2012 and beyond will bring 40 and 100GbE products. These technologies will continue to co-exist and EMC will continue to deliver solutions to address customer challenges and requirements.
Direct-attached storage architecture has distinct limitations. Only one server can use the storage attached to it. Storage that goes unused becomes a “stranded” or underused resource. If an application is saturating one server, you cannot deploy another server to share the burden because the application data is only available on the direct-attached server.
Note to Presenter: View in Slide Show mode for animation. Deploying a SAN lets you share storage. By consolidating with a SAN, you can reduce your total cost of ownership and eliminate needless duplication and resynchronization of files. SANs provide greater storage usage through consolidation. You can easily allocate storage to servers as needed. And, as we’ll see on the next slide, a SAN makes backup easier and more efficient.
Note to Presenter: View in Slide Show mode for animation. In a traditional backup scenario, data is moved from the server to a tape drive. While backup is occurring, most applications are offline, unavailable to users. This impacts revenue—as well as the LAN. In a SAN, backups can be performed on cloned copies of data, so that production environments are not interrupted. SANs also shorten the time it takes to restore lost data. In a SAN solution, the recovery of lost information can be significantly accelerated through the use of a cloned copy. Instead of having to wait for tape, the cloned copy can be mounted in minutes to the failed server. In addition, you can keep copies of the data from multiple days online. And in the event of a hidden corruption, you can recover faster—just by mounting an initial copy; checking it for corruption; and, if corruption is found, mounting a previous copy—all in a matter of minutes. Compare that process with recovering from tape.
More and more customers are telling EMC that tape-based recovery in the event of an unplanned outage doesn’t work. Restoring from tape after retrieving the tapes from a remote site is no longer viable: It simply takes too long, and there are no guarantees that you will be able to recover all your mission-critical information. So, if you do backups at different points throughout the night, and a disaster is declared at 10:00 a.m., how do you go back to a known good point in time at your remote site? You can take the existing investment you’ve made in a Fibre Channel SAN and, when you are ready to conduct disaster recovery, simply add some remote-replication software like the EMC SRDF family or MirrorView, and put a second storage array at your remote site. That’s all you need to do to ensure that all your mission-critical information—down to the last transaction—has been captured. And you can be up and running in hours, as opposed to days.
With VPLEX, you can achieve true High Availability – at distances that were not possible before. This means operations continue and data remains online even if a failure occurs. No manual intervention is required, no recovery or restart is necessary. In fact, within synchronous distances using VPLEX Metro, you can think of VPLEX as providing Disaster Avoidance instead of just Disaster Recovery – because there is literally no disaster event from which recovery is necessary. The benefits to high availability from a VPLEX solution are substantial: Performance is improved through lower latencies tied to distance, as well as the caching architecture of VPLEX Fault domains are truly isolated. No concerns around a device inadvertently impacting other devices in the network True Zero RTO means exactly what it says – solutions can be architected without a requirement for “recovery” Ease of management … When two VPLEX clusters are connected together with VPLEX Metro or VPLEX Geo, VPLEX gives customers shared data access between sites, meaning the same data, not a copy but the same data, exists at more than one location simultaneously. Instead of using costly and complex processes, and having to choose RPOs and RTOs, VPLEX can withstand a component failure, a site failure or loss of communication between sites and still keep the application and data online and available. This is true High Availability, with VPLEX deployed in a real active-active configuration. The architecture is further improved through a new VPLEX Witness. The VPLEX Witness exists in a different failure domain from either VPLEX cluster to arbitrate between sites, if necessary, when a failure occurs. With VPLEX, users can build an architecture that is 100% non-disruptive for all hardware and software upgrades. So with VPLEX, you can provide a solution that allows instant access to information in real-time, eliminates operational overhead and reduces complexity, and reduces the time required to copy and distribute data across locations.
Fabric-based encryption is appropriate for any company who must comply with common security standards including PCI DSS and Sarbanes-Oxley or needs to protect data at rest on disk that is routinely transferred offsite physically or over the wire. It offers a number of advantages to customers with SANs who wish to secure their data at rest: It has the lowest total cost of ownership compared to other hardware-based encryption methods due to its minimal impact on the SAN and its heterogeneous support. It plugs into your existing SAN without disruption or downtime, and uses the same administrative tools as non-encryption switches for unified management. It has minimal impact on system performance due to its split fabric architecture. It scales much better than appliance- or tape-based encryption solutions and eliminates the labor-intensive process of identifying and classifying confidential information stored throughout the enterprise. It also is supported by RSA Key Manager for the Datacenter, the industry-leading key management system.
Benefits of purchasing SAN products from EMC
Info on FC growth: http://www.infonetics.com/pr/2010/4Q09-SAN-Equipment-Market-Highlights.asp FC to grow over next five years: http://www.reliant-technology.com/storage_blog/fibre-channel-and-brocade-dcx
New 10 Gigabit Ethernet Conveged Storage Networking switches support multiple protocols – FCoE, iSCSI, NAS. Some converged network switch models support Fibre Channel to connect into an existing Fibre channel environment. FCoE supports Fibre Channel natively over Ethernet. In this picture we see virtualized servers with Converged Network Adapters, or CNAs. CNAs replace both HBAs and NICs. Today, CNAs attach to a converged network switch, which provides connectivity to the corporate LAN, NAS and iSCSI storage, native FCoE storage, and/or traditional Fibre Channel storage. Reducing the number of adapters saves the number of cables required and with fewer adapters, power is also saved. FCoE, iSCSI and NAS protocols not only use the same adapter, they also use the same cable. Also, cable bulk can impede airflow which may cause increased power consumption. Finally, new in-rack copper cables reduce power consumption but are robust enough to carry storage traffic. Cabling and adapters cost savings are a significant component of I/O consolidation cost savings. As shown in the slide, EMC’s storage products (VMAX, VNX, Isilon and CLARiiON) support multiple protocols. As far as what connectivity to choose, a customer’s decision process begins with the retrieval method of their stored data – Block or File? Blocks are the most elemental units of storage and refer to the largest amount of data that a single operation can access. Storage Area Networks (SANs) access blocks directly. SANs utilize Fibre Channel, FCoE or iSCSI protocols. Files consist of several blocks. In Network Attached Storage (NAS) systems, the server organizes blocks into files. NAS utilizes the IP (Ethernet-based) network to move files using CIFS or NFS protocols. Source Note: Minimum 8% power reduction comes from Cisco TCO whitepaper.
The case for transformation is really rooted in our legacy – in how we’ve behaved, how we’ve been run, and how we’ve been organized in the past IT-as-a-Service means taking that virtual hosting platform and layering on top of that some key capabilities involving extensive automation The real goal is “optimizing IT production for business consumption” – what’s good for IT may not always be what’s good for the business If we are going to compete with external IT service providers, we have to be organized and operate like IT service providers that can win the business A New IT Business Model Now we will focus on transforming to a model that looks more like a professional services organization than a customer services organization We have to start with the services as opposed to the technology – our clients are not consuming a technology, they are consuming a “business capability” Enabling Technology This is probably the easiest of the three, and the main focus of our first two phases DNA, Skills, Roles & Organizational Alignment – this is our focus now Important to focus on “front office” capabilities – “front office” meaning the marketing and sale of IT services
Virtualization technology is causing significant changes across storage networks. Most significantly, server virtualization technology is driving demand for networked storage solutions due to the overall increase in storage capacity requirements brought about by virtualization initiatives. To fully realize the benefits of server virtualization ( e.g., mobility, workload balancing, increased utilization, and high availability), businesses will need to deploy SAN infrastructures to consolidate and share storage resources. Users are therefore purchasing additional storage capacity and networked storage technologies to satisfy the storage requirements associated with server virtualization deployments. Increasing use of server virtualization is also leading users to rethink data protection and disaster recovery strategies, which is also expanding the market for remote data replication solutions. These are the top reasons for deploying storage area networks with server virtualization: Mobility of virtual machines between physical servers Easier and more cost-effective disaster recovery Increased uptime and availability Simplifies maintenance and upgrade of physical server Ability to store multiple copies of virtual machines for high availability 72 percent of virtual servers are SAN-attached
Over short distances such as within a data center, SANs are typically extended over optical links with multi-mode Optical fiber. As the distance increases, such as within a large data center or across a campus, single-mode fiber or single-mode fiber with Course Wavelength Division Multiplexing (CWDM) is typical. Across metropolitan distances, Dense Wave Multiplexing (DWDM) is preferred. Fibre Channel over IP (FCIP) can be used to extend a Fibre Channel SAN across any distance. FCIP can be used across metropolitan and campus distances—or across intercontinental distances where IP might be the only transport available. SAN Extension over FCIP typically has many cost benefits over other SAN-extension technologies, as it is relatively common to have existing IP infrastructure between data centers that can be leveraged at no additional cost. FCIP tunnels Fibre Channel frames over an IP link, using TCP to provide a reliable transport stream with a guarantee of in-order delivery. iFCP (Internet Fibre Channel protocol) is a gateway-to-gateway protocol for providing Fibre Channel fabric services to Fibre Channel end devices over a TCP/IP network, like FCIP or iFCP.
Two emerging network standards can address the problem of multiple networks: Fibre Channel over Ethernet is a storage networking protocol that supports Fibre Channel natively over Ethernet. Fibre Channel over Ethernet encapsulates Fibre Channel frames into Ethernet frames, allowing them to run alongside traditional Internet Protocol (IP) traffic. Today’s data centers use Ethernet for TCP/IP networks and Fibre Channel for storage area networks. With Fibre Channel over Ethernet, Fibre Channel becomes another network protocol running on Ethernet, alongside traditional IP traffic. This means Fibre Channel over Ethernet runs alongside IP traffic on Ethernet, unlike iSCSI which runs on top of IP using TCP. Fibre Channel over Ethernet is designed for any Ethernet speed, but another emerging protocol, called Data Center Bridging (DCB), is required for performance and reliability. The Data Center Bridging standard makes Ethernet as reliable for storage traffic as Fibre Channel because it creates an Ethernet that drops frames as rarely as Fibre Channel. Note to Presenter: Cisco refers to the new Data Center Bridging standard as “Cisco Data Center Ethernet.”, and it is also reffered to as Converged Enhanced Ethernet (CEE) Fibre Channel over Ethernet is a storage networking protocol that maps Fibre Channel natively over Ethernet. It… Was developed by International Committee for Information Technology Standards (INCITS) T11 Fibre Channel Interfaces Technical Committee Allows native Fibre Channel to run on Ethernet alongside traditional IP traffic Seamlessly integrates with existing Fibre Channel networks, management processes, and workflows Fibre Channel over Ethernet requires three modifications to Ethernet in order to deliver the capabilities of Fibre Channel in SANs: Encapsulation of a native Fibre Channel frame into an Ethernet frame Extensions to the Ethernet protocol itself to enable a “lossless” Ethernet fabric Replacing the Fibre Channel link with MAC (Media Access Control) addresses in a lossless Ethernet Converged Enhanced Ethernet is being developed by the IEEE Data Center Bridging Task Group. The new enhancements to traditional Ethernet make Ethernet a viable transport for storage traffic. PAUSE and Priority Flow Control Enhanced Transmission Selection Data Center Bridging Exchange Notification