by clicking on the page. A slider will appear, allowing you to adjust your zoom level. Return to the original size by clicking on the page again.
the page around when zoomed in by dragging it.
the zoom using the slider on the top right.
by clicking on the zoomed-in page.
by entering text in the search field and click on "In This Issue" or "All Issues" to search the current issue or the archive of back issues respectively.
by clicking on thumbnails to select pages, and then press the print button.
this publication and page.
displays a table of sections with thumbnails and descriptions.
displays thumbnails of every page in the issue. Click on a page to jump.
allows you to browse through every available issue.
FCW : May 30, 2014
39 at preventing network loops that bring communication to a standstill. But this comes at a trade-off because the network will also regularly hold frames, slowing performance. But for many enterprise networks, blocking switches deliver acceptable throughput and remain workhorse devices for many apps. When high performance is required, such as for top-of-rack switches, IT managers may opt for pricier devices using nonblocking protocols, such as Transparent Interconnection of Lots of Links (TRILL), to address loop and performance issues. As for routers, their main job remains forwarding Layer 3 IP addresses to an appropriate IP subnet. Many organizations increasingly are using hybrid devices (such as routing and IP switches) that can operate on both Layer 2 and Layer 3 of enterprise networks and combine the packet-handling capabilities of routers with the speed of switches. The performance of this class of switches is aimed at the core or distribution layers in large infrastructures. Intelligence in these multilayer switches identifies common traffic flows and switches those flows on the hardware-embedded, application- specific integrated circuits. For traffic outside the normal flows, the multilayer switches use routing functions. This enables the higher overhead of routing. Looking Ahead to SDN and Fiber Traditional network switches and routers typically come with both a data plane (the part of the device that forwards packets in one interface and out another) and a control plane (a component that builds the rules for forwarding packets). Now, some network managers are investigating an alternative technology. Software- defined networking moves the network control and decision-making responsibilities away from the switching devices and into a separate network element: the flow controller. Through SDN, network makers aim to simplify network management, enable network reconfiguration and enhance understanding of what's actually happening on the network. Researchers are still studying whether SDN can be used to build better, and more flexible networks. But for now, it's a wait-and-see approach. "The guidance we are giving our clients is to make sure the new network solutions they're buying are SDN-ready, but they shouldn't rush to deploy SDN functionality for architectures in production today," says Jason Nolet, vice president, data center switching and routing, at Brocade Communications Systems. "Over time, SDN will demonstrate that it has value, but that time is not here yet." As new apps continue to fuel demands for increased capacities and throughput in enterprise networks, IT managers also can look beyond switches and routers to the pipelines that interconnect the infrastructure. For some organizations, that means expanding the use of fiber-based communication links. But modernization roadmaps looking further than three or four years must account for a fundamental change in fiber technology, Fratto notes. "If you're planning to go to 40 gigabits at some point in the near future, your current fiber plant probably won't support that high throughput," he says. "Today, you can cut fiber and attach the ends in the field, but when you go to 40 gig and beyond, field termination isn't possible because all that fiber is custom made." His advice: Work with a trusted network provider who can define the necessary lengths of fiber strands for each termination point and then build the proper links in a clean room. "If you have fiber that supports 1 gig or even 10 gig, you're OK for now," Fratto says. "But if you're putting in new fiber, talk to your installer about how to transition from a 10-gigabit resource to 40 gigabits and beyond." architecture copies the entire frame, latency varies with frame length. A speedier alternative is a cut-through switch, which only looks at destination addresses before forwarding frames. Of course, bypassing CRC means some packet errors will slide by undetected and may propagate throughout the network. For this reason, many network managers are balancing data accuracy and speed with switches that support both cut-through and store-and-forward architectures. To take advantage of the latest switches and boost network performance, some IT managers are focusing on top-of-rack switches. e goal is to create higher port densities by moving up from 48-port, single-unit switches to 96-port, two-unit configurations. In addition to packing in more ports, this approach also boosts uplink capacities. "It's at the top of the racks where some organizations are moving from standard 10-gigabit to 40-gigabit uplinks for data centers that need that higher capacity," says Mike Fratto, principal analyst at the market research firm Current Analysis. e right protocols can also ensure that switches keep network traffic flowing efficiently. Switch protocols fall into two basic categories: blocking and nonblocking. ese protocols come into play when switches have two or more connections to another switch. If the switch has a shared bus or the switching components cannot handle the theoretical total of all ports on the switch, it is considered a blocking switch. "If a blocking switch has 48 1-gigabit- per-second ports and can process only 24 gigabits at a time, it's going to run fine until it gets to about 24 gigs and then it will have to hold frames until capacity is freed up," Fratto explains. "With nonblocking switches, every frame that comes in can be processed on the way out, so there's no oversubscription of the switch." Blocking protocols, such as the widely deployed Spanning Tree standard, have proven to be efficient CDWG.com | 800.808.4239
May 15, 2014
June 30, 2014