 | Issue: | Asia-Pacific II 2007 |
| Article no.: | 12 | |
| Topic: | Data centre strategies | |
| Author: | Steven J. Russ | |
| Title: | VP of Global Marketing Programs | |
| Organisation: | Avocent Corp. | |
| PDF size: | 256KB |
About author
Steven J. Russ is the VP of Global Marketing Programs at Avocent Corp. Mr Russ has over 15 years of operating experience in enterprise IT markets with both public companies and venture-backed start-ups. Prior to joining Avocent, he served as executive-in-residence at Corporate IP Ventures. He has held senior sales, marketing, business development and consulting positions at Unisys, Dun & Bradstreet, The Sentry Group (acquired by Gartner Group), PaySys (acquired by First Data Group), Internet Security Systems and Trusted Network Technologies. He is the author of numerous articles on marketing and customer value. Mr Russ holds an MBA in finance and marketing from Columbia University’s Graduate School of Business, where he was a Unisys scholar. Additionally, he is a meritorious graduate of the US Naval Academy with a Bachelor’s degree in systems engineering. He also holds a Master of Science equivalent in nuclear engineering from the US Naval Nuclear Power Program.
Article abstract
Managing scattered, remote IT assets can be a big challenge. As businesses go global, IT assets go global as well. By using remote management tools the cost of preventing or repairing outages goes down dramatically, as does the time needed to respond to systems failures and repair them. Separate, ‘out of band’ communications channels give the support staff control over equipment halfway around the world; they can monitor it and often correct problems before they seriously impact the operation.
Full Article
Over the years, the role of IT staff has changed almost as much as the underlying technology itself. While they’ve always been the ‘go-to’ guys for internal technical problems, demands on this valuable team have grown as corporate computing has become more distributed. The extended enterprise – customers, partners, suppliers, remote employees – depends more and more on an always-connected, always-on IT infrastructure. Given the all-important role infrastructure performance plays in business success, IT professionals have had to add ‘business specialist’ duties to their technology related obligations. IT’s growing importance to the business has not spared it from the ongoing pressure of global competition, which makes it necessary to cut costs while simultaneously improving service and productivity. Infrastructure must be available 24/7 while efficiency constantly improves and costs are reduced. The answer: while there is no single silver bullet, remote management strategies and tools are a key piece of the puzzle. Typical IT infrastructure The typical IT infrastructure of today’s large enterprise consists of one or more data centres and may also include several remote sites or branch offices. In some cases, large enterprises deploy redundant data centres to ensure business continuity and consistent IT service availability in the event of a disaster. Architecturally, data centres share a similar makeup that includes servers, storage and applications, and a network infrastructure of hubs, routers, firewalls and switches. Whether an enterprise’s infrastructure is large and dispersed across multiple sites, or limited to a single data centre, its IT managers all face similar challenges – to deliver the highest level of performance and availability at minimal cost. Typically, large organizations depend on complex systems-management frameworks (such as HP OpenView, IBM Tivoli, CA Unicenter, BMC PATROL) to monitor network performance and manage application performance and availability, while small- and medium-sized enterprises may opt for more focused, less expensive applications. These management tools depend on the IT network to monitor performance and productivity, and perform effectively as long as network connections remain available. Remote management with out-of-band tools Whatever the size or makeup, remote management is critical; it saves time, which saves money, so it’s been readily embraced by IT managers. Remote management allows the administrator to access and control IT assets through a variety of networks, serial or modem connections via the network infrastructure, or via a path separate from the network infrastructure – out-of-band – from a distance of 20 feet to thousands of miles away. Out-of-band management refers to the use of a dedicated ‘management’ channel to monitor and maintain devices. In-band management, on the other hand, uses regular data channels for these functions. Out-of-band management tools provide secure, alternate paths to access remotely, monitor and manage IT assets throughout the network infrastructure. Out-of-band management is a best practice when network connections are unavailable. If an IT asset, or the network, goes down, out-of-band management tools can restore the asset in the shortest time possible. Out-of-band management tools also minimize the need for local management and site visits, dramatically reducing the time and operational costs needed to bring IT assets back online. Out-of-band tools are increasingly being used ‘in band’ as the bandwidth of the in-band network becomes increasingly valuable. The relationship between the network infrastructure and out-of-band management tools is typically configured in a manner similar to the topology shown in this first illustration. Let’s look at a scenario to show how out-of-band management tools may be used. In this illustration, a server in the data centre goes down but the network infrastructure remains operational. Using out-of-band management tools and the network infrastructure, the administrator accesses the IT asset, diagnoses the issue and restores it to the network. There, it can be managed back to productivity using a systems management application. The benefits of using out-of-band management tools include lower labour costs, increased productivity and reduced risk. Operational costs are decreased and IT asset availability is increased. Ideally, out-of-band management tools should support industry-standard security – including authentication, directory and encryption protocols – to enable integration with existing security infrastructures. For IT managers tasked with increasing productivity and reducing costs, out-of-band management tools are ideal for supporting next-generation IT infrastructure. Out-of-band management tools So what exactly are out-of-band management tools? They are software, hardware, embedded technologies, or a combination thereof. Out-of-band management software provides consolidated access, change management and configuration management for tools such as serial console servers, KVM switches, power management appliances and service processor managers. It also provides the ability to manage diverse IT assets connected to these out-of-band tools by way of a single consolidated view, while providing the scalability needed to meet the demands of the largest enterprises. Serial console servers provide remote access to the serial management ports included on some servers and other network IT assets (routers, switches, cables, firewalls, etc.) rather than depending on the network connection, essentially to keep IT assets up running when the network is down. KVM, keyboard, video and mouse, switches or KVM-over-IP switches access servers through ports to provide access as if the administrator were physically present, enabling remote IT staff to control servers easily and efficiently from a distance. Intelligent power distribution units, IPDUs, provide the ability to power on and off remote equipment for operational control or recovery from software/hardware failures. Service processor managers provide consolidated, centralized access to embedded service processors. A service processor is a separate, dedicated internal processor located either on the motherboard of a server, a PCI card, on the chassis of a blade server or on the telecommunications platform. Since it operates independently from the server’s CPU and operating system, OS, it helps IT staff gain control when the CPU or OS is locked up or otherwise inaccessible. Service processor managers allow administrators to reboot servers whether or not the main processor or the OS is operational. Intelligent Platform Management Interface, IPMI, HP Integrated Lights Out, iLO, and Sun Advanced Lights Out Management, ALOM, are examples of service processor technologies. These service processors are one of the newest components in server management strategies. While the first service processor technologies and related protocols were proprietary, more recently Intel, HP, Dell, IBM and other hardware vendors collaborated to develop the open-standard service processor called IPMI that is now included on many rack-mounted and blade servers built on Intel’s X86 architecture. In late 2004, the first vendor-independent IPMI manager emerged. Just last year, the first vendor-independent service processor manager came on the market. Research by industry analyst firm Gartner shows the number of servers shipping per quarter is still on the rise, even with the growing adoption rate of virtualized servers. Since service processor technology exists on most servers purchased over the past three years, it makes economic sense to leverage technology enterprises already owned. There are two ways to access service processors: either a dedicated Ethernet interface (out-of-band) or a shared data Ethernet interface (sideband). Notice in the illustration below that in the sideband case, although there is only one Ethernet jack in use at the back of the server, there are still two Ethernet hardware sets inside the server, which means those interfaces can have separate IP addresses. While the feature set of service processors varies depending upon the service processor type, all provide remote power control (on/off/cycle/status) and remote console access via Serial over LAN, SoL. Some also provide server health monitoring (fans speed and status, temperature, voltages), graceful, (OS-level) shutdown capabilities, remote KVM and Virtual Media capabilities. Return on Investment The cost savings of using out-of-band tools is significant. A European telecommunications company used out-of-band management tools to deploy more than 2,000 additional IT assets within an existing IT infrastructure over a two-year period – all without additional support staff. Operational costs were reduced as its overtime labour costs decreased by 92 per cent. Time needed to deploy assets dropped by 50 per cent, enabling the company to win in competitive bidding situations because competitors could not match its deployment speed. Personnel productivity increases enabled the company to expand IT assets by 33 per cent per year to more than 100 sites, without the need to hire additional staff. Within 16 months, the out-of-band management tools paid for themselves. Deploying out-of-band management tools was a cost-effective, efficient and smart business decision. Keeping IT infrastructure ‘always on’ is people, time and money intensive. Remote or out-of-band management provides a better way – a time-saving, secure and cost-effective method – to ensure that next-generation IT assets remain productive and connected to the network.
