TCO and ROI are a part of the justification for SANs, however for most organizations the precise final analysis is measured in additional-tangible advantages.
by Alan R. Earls
The U.S. Air drive's forty fifth house Wing, headquartered at Patrick Air force Base close Cocoa seashore, FL, exemplifies a virtually literal "mission-essential" storage area network (SAN) consumer. It helps and manages house launch classes for the department of defense, NASA, and commercial entities, and generates tremendous quantities of facts having to do with space launches, all of which ought to be backed up or saved to protect in opposition t statistics loss.
"The probability of hurricanes and flooding on the Florida coast makes it completely critical that they back up their facts," says Glenn Exline, supervisor of superior technologies on the area Wing. "With their previous storage system, backup took as much as 14 hours each day. With their SANs, they have decreased that to best two hours a day."
The Wing's SAN success also could be considered an argument that storage networks can be justified via something apart from a return on investment (ROI) examine. The corporation's SAN installing now not most effective helped dramatically reduce the backup window, however additionally helped to consolidate a network from 52 servers to 35 servers, simplifying server administration. in addition, with fewer servers on its community, the Wing expects to spend less going ahead to refresh its servers as a part of the alternative cycle.
while some agencies have managed to position collectively ROI and TCO stories to support the argument for SANs, other organizations locate that bean counting is never worth the trouble. instead, they cite clear productiveness features, new alternatives, more desirable reliability, and different merits that they say make SAN economics a no brainer.
based on John Webster, founder and senior analyst on the statistics Mobility group consulting company, when a SAN undertaking is within the offing, "each person starts showing up with an opinion—or a priority." regularly, people are on the protective, involved that trade of any type may have a negative influence on how they do their job. "they arrive to the desk saying, 'convince me.' lots of those americans may be unmoved with the aid of any form of ROI argument. I inform providers they should aid their valued clientele allay the fears of internal customers, principally corporate security officers and database administrators, if they are going to sell SANs efficaciously," Webster says.Multi-step method
An illustration of the need to sell internally—and the capabilities results—is provided through Bob Massengill, manager of technical capabilities at the Wake wooded area university Baptist clinical center (WFUBMC), a clinical middle and educating sanatorium in Winston-Salem, NC. Massengill's organization applied its first SAN in 2000, essentially to address two problems: an e mail gadget that saved littered with disk screw ups, and a desire to spend prudently when the mainframe disk systems went off lease. "i was faced with the alternative of staying with mainframe-best storage or going to a SAN and also essential to do whatever about managing the disks on a whole lot of servers," Massengill explains.
maintaining everything up and running had turn into very labor-intensive. What's greater, although lots of the servers, principally those operating home windows NT, were configured similarly, "we saved operating into the problem of getting midway via a challenge on a selected server and running out of disk space." The simplest purposeful answer changed into to improve that server with extra disk storage. in the meantime, though, a number of different servers had scarcely tapped their accessible storage—that means that average means utilization become very low.
but getting purchase-in on the SAN from the a lot of departments that felt that they had ownership over certain servers wasn't convenient, recollects Massengill. He bought the initial part of the implementation as a "proof of conception" developed round an EMC Symmetrix array, which became at the start deploy to support simply e mail purposes and the mainframe. "At that factor they felt it might finally grow incredibly, however from a funding standpoint they could not bank on it," says Massengill.
He needn't have concerned. through late 2000, with the Symmetrix entirely mirrored, disk screw ups stopped, mainframe I/O waits disappeared, and americans retrieving ancient e-mails had been "blown away" by way of the more desirable performance. "notice unfold fast and relatively quickly they had departments lining up and asserting they desired to return aboard," remembers Massengill. What he did need to be troubled about, in fact, became being able to develop fast enough. "We began placing more users on the Symmetrix and populating their sixteen-port McData switches, and inside six months they had outgrown them," he recalls. That led him to improve to director-category switches from McData. "We put in a sixty four-port Intrepid and stored both sixteen-port switches in vicinity," he says. through then, storage on the SAN had rocketed from just over 2TB to about 8TB. It now stands at 45TB.
but the SAN had no longer yet solved all of Massengill's issues. For the health center, delivering the maximum stage of patient care and clinical schooling depends on safeguarding facts and guaranteeing that sanatorium tips systems are extremely attainable. additional urgency changed into provided via the increasing protection and facts retention necessities mandated by using HIPAA.
"A key situation for us became records integrity," explains Massengill. "The medical middle turned into experiencing repeated media disasters and corrupted facts on tapes. because the sanatorium become leasing the tape drives, it determined to substitute them with StorageTek T9940s when the leases expired early this year."
issues about skill and centralization of backup operations brought about the IT department to seek a new answer. The scientific center crucial further storage capacity for its Unix and NT servers. They additionally wanted to centralize backup operations in an environment that blanketed an IBM zSeries 800 mainframe, a Hewlett-Packard NonStop Himalaya S7000 server, and a network of Unix and NT servers.
The mainframe residences the patient accounting gadget, while patient remedy facts reside on the Himalaya. The scientific middle uses the Unix servers for PeopleSoft monetary and human substances functions, net, and intranet applications. in line with Massengill, the desires of the improved SAN project were to boost protection and integrity for scientific information facts, affected person accounting records, and health center administrative facts; reduce the variety of hours vital for day by day backups; reduce the number of tapes necessary for backups; and free the IT body of workers from tape mounting chores.
however, notes Massengill, the SAN wasn't meant to quite simply resolve these instant wants. As a part of the brand new SAN, a StorageTek virtual Storage manager (VSM) system become delivered "to aid us maneuver as the inevitable adjustments happen down the road," he says.
The extended SAN changed into carried out early this year. "In October 2002, they brought the open systems servers over to StorageTek PowderHorn tape library silo, and in February 2003, they implemented StorageTek VSM and the mainframe piece going into the STK PowderHorn."
in response to Massengill, incorporating the tape library in the SAN led to a lessen in tape count number from about 30,000 to simply 600. It also saved ground area and tape mount time. before the project, MVS operators were mounting 1,000 to 1,200 tapes day after day, he says. Mount instances were approximately three minutes. "these days the workforce has considerably fewer tapes to tackle and mount instances are in the sub-2d latitude," he says. "That means IT staff has time for equipment monitoring, proactive protection, and practising."
Integrating the tape libraries and SAN also tremendously diminished media disasters and diminished backup time by three hours per day. "in the past," says Massengill, "batch cycles ran except 6 AM, so users were unable to access the equipment until these jobs have been finished." youngsters, hospitals begin admitting patients for day surgery at 6 AM, so if jobs ran beyond that hour it impacted the admissions and medical staff. "today those batch jobs are complete through three AM, offering greater system entry to users," he says.efficiency is vital
Becky Davidson, records supervisor for EDS/Sara Lee Bakery group, in St. Louis, faced fewer complexities in justifying a SAN, however subsequently found identical proof points. She says the SAN changed into applied to address a couple of issues, principally the deserve to enhance the number of connections into the enterprise's IBM commercial enterprise Storage Server ("Shark"). "We wanted to have twin paths and they desired to connect a huge number of machines," says Davidson. Minimizing the cabling hassles linked to SCSI become additionally a riding force behind the SAN. Yet an extra aim turned into expanded performance.
"We did not have lots issue justifying the SAN both in the account or within EDS," says Davidson. partly that changed into as a result of Sara Lee is turning out to be unexpectedly so it become handy to show that and not using a SAN they would run out of attachable disk space. "Our customer become additionally very concerned about efficiency, and they could show that multi-pathed Fibre Channel changed into faster than SSA or multi-pathed SCSI," says Davidson. The EDS team turned into also able to demonstrate that there became a greater reductions per GB with a SAN.
identifying a SAN dealer turned into also effortless. "We run IBM servers connected to IBM Shark and IBM tape drives so they went with IBM," she says. in the beginning, they all started out with three sixteen-port 1Gbps switches from Brocade—one for development and two for construction. As they grew they delivered a fourth sixteen-port switch for backup site visitors. consequently, they delivered three area switches to join a 21-drive tape library. Altogether, they now have two 32-port 2Gbps switches, three 8-port 1Gbps switches, and the 4 sixteen-port 1Gbps switches.
"The largest improvement was flexibility," says Davidson. "If they wish to add an extra system they can just plug it right into a switch, zone it, and we're decent to move."
On the downside, Davidson says that zoning can also be complex and the material conception can also be a little overwhelming. happily, Brocade's management GUI lets Davidson verify things out on the switches although the facts middle is determined in other places "so i will tell what's plugged in, what is active, what's having problems, and what isn't working."
Davidson says that the SAN solved a lot of problems, "nonetheless it does add more things to preserve song of."SAN-based mostly replication
even more upbeat in his SAN evaluation is Gary Pilafas, a senior storage and techniques architect at United Loyalty services, a unit of UAL Corp. (the dad or mum of United airlines), in Elk Grove Township, IL. United Loyalty features operates an OLTP engine that requires information replication.
Pilafas uses Hitachi's Shadow image to do element-in-time copies from one facts core to another. With the SAN, says Pilafas, it's viable to duplicate and then run queries against one server and reviews in opposition t one other. "You cannot do that in a DAS [direct-attached storage] ambiance with out lots of ache," he says. Replication over an IP network is completed with the Fibre Channel over IP (FCIP) protocol.
In Pilafas' case, justifying the SAN turned into easy because the company had been locked into an expensive internet hosting association with a storage features issuer, so an entire ROI analysis took place very straight away.
Between 2001 and 2003, United Loyalty services' SAN grew from 4TB to 25TB.
Pilafas says the SAN has additionally proven its worth since it provides high availability, which could support enterprise continuance and business resumption capabilities. And the SAN also helps the united.com earnings engine—the e-commerce entrance-end for UAL Corp.no longer so elementary
but from time to time even superb results may also be coloured by using doubts. Xcel power, in Denver, developed a plan to implement a SAN in 2000. on the time the task began, "we concept they vital extra things than perhaps they did," says David Czech, a storage programs engineer at Xcel. Xcel's SAN is in accordance with a Hitachi statistics systems Freedom 9960 array, McData switches, and a StorageTek tape library.
Xcel's SAN dreams had been to consolidate storage, simplify administration, and boost capacity utilization. moreover, the business wanted to be capable of without problems replicate to yet another facility and to share tape gadgets.
"Some of these issues came about and a few did not," admits Czech. while the SAN did make it possible to consolidate to fewer storage contraptions, nothing got simpler. "I don't suppose it made the rest less difficult to control, and now there are lots of additional steps concerned in allocating and troubleshooting issues that are attached to the SAN," he says.
still, Czech characterizes the problems as generally annoyances. And he is now taking a look at application equipment to help with SAN administration and storage useful resource administration (SRM). longer term, Czech says he's searching on the standard company storage infrastructure and the skills for virtualization. he is also searching at the potential for consolidating RAID arrays within the SAN.
Richard Boud, technical account supervisor with Technica UK Ltd., an unbiased storage consultancy in the UK, sees large and consistent move toward SANs amongst his valued clientele. He says the fundamental benefits that shoppers see as drivers for imposing SANs are right here: flexibility in storage allocation; adapting to changing business necessities; improved performance, primarily backup enhancements (i.e., snapshots) and catastrophe healing (i.e., faraway replication); better statistics availability; centralization, mainly for ordinary services such as backup; and simplified storage ..management.
Boud says that ROI and TCO measurements "are pointed out an awful lot, however likely more through providers than anyone else. Most consumers take ROI and TCO claims with a large pinch of salt. They shouldn't have the time, tools, or correct statistics to consider skills or exact ROI. in its place, they opt to be introduced with credible justifications and to make use of standard experience to tell them the place know-how can essentially enrich their infrastructure."
Alan R. Earls is a freelance writer in Franklin, MA.Consolidating SAN 'islands'
by using Robert Strechay
The transforming into need to consolidate remoted storage enviornment networks (SANs) stems partly from the indisputable fact that SANs had been configured in an ad hoc fashion to help selected purposes. These SAN "islands" sufficed for a long time, but then applications were prolonged and storage necessities increased.
on the equal time, IT professionals' mandate was to "do greater with less." Even because the variety of SAN islands swelled with greater site visitors and data, fewer dollars have been obtainable to appoint experts to re-architect the SANs.
as a result, groups are recognizing that the simplest approach to grow while reaching return on investment (ROI) is to consolidate their statistics-core components and applications, increasing the density and scalability of their storage networks. however this requires careful planning of future storage networks so one can scale past the island idea.
The emphasis agencies place on business approaches to reduce downtime, administration complexity, and value varies. These elements are often not a part of the SAN design system, yet they may still be.
despite application and business needs, most IT experts are busier than ever, because of the financial downturn. Layoffs and downsizing have despatched personnel scrambling to support further and further enterprise devices, yet applications continue to roll out and facts accumulates hourly.
as a result of a shortage of components, many IT departments took shortcuts with their early SAN deployments. They counted on efficiency features devoid of further management overhead and troubleshooting. When some SANs failed to operate smartly, or site visitors grew too impulsively, departments regularly connected their personal direct-connected storage (DAS). If the functions using DAS outperformed the SAN-connected functions, the enterprise gadgets would demand that company IT revert lower back to supporting DAS for his or her purposes.
The purposes doubtless would have labored excellent on the SAN if it had been accurately designed. sadly, few IT stores had the essential knowledge. And the style SANs had been marketed did not aid either. SAN gadget became frequently sold via OEMs—usually the principal gadget carriers. in this context, SAN machine was a sliver of the universal purchase. With companies and customers giving little thought to future increase of the storage network infrastructure, no wonder many early SAN installations disenchanted.
Yet, many different OEM-supplied SANs labored exceptional. They met many shoppers' deserve to support utility deployments and to inexpensively leverage existing storage arrays. The clients also could leverage their OEM carrier contract to get support with deployment and ongoing SAN administration.
Storage community upkeep—starting from executing moves, adds, and changes to adding or combining logical unit numbers (LUNs)—frequently requires a provider name and an on-web page seek advice from from an o.e.m technician. This OEM dependence reduces the IT department's visibility into storage community considerations.
Outgrowing a SAN usually capacity adding an extra island, or becoming the present SAN with additional switches and inter-change hyperlinks (ISLs), which will also be complicated and cause interoperability and performance problems. Unpredictable ISL efficiency has partially led to the deployment of remoted SAN islands.
furthermore, SAN management is 5 to 10 years behind common LAN and WAN network management. standard LAN administration features, equivalent to notification of congestion or different faults, are not extensively accessible on SAN switches.
although SAN islands have proven value, many companies are outgrowing them. while they add more and more SAN islands, they reduce the ROI of the presently deployed SAN islands as a result of the better charge of management. it's time to devise for strategic, lengthy-time period storage networks.Political impediments
Getting everybody on the identical SAN consolidation group is complex due to body of workers shortages and political issues. Few gifted storage specialists exist: there's possibly one storage architect for each 10 LAN/WAN network architects. This reflects the undeniable fact that storage isn't considered as part of the community aspect of IT but, fairly, as an outgrowth of the purposes aspect.
Storage has transitioned into the server administrator's area. but in most situations, server administrators are too busy to learn satisfactory about storage networks. Compounding concerns is the longstanding distrust between server administrators and traditional community administrators (neither neighborhood thinks the other is responsive satisfactory), which further complicates SAN consolidation projects.
regardless of these difficulties, some companies are starting to respond to the fallout from SAN island growth through looking at their storage networks the equal means they do their LANs and WANs. The political intent is evident: company gadgets are chuffed with the LAN and WAN services, so let's deal with SANs the identical method. IT corporations additionally appreciate the reliability, predictability, scalability, safety, and visibility they benefit through having a unified architecture, as opposed to an advert hoc assortment of separate networks.Put the "N" back in SANs
modern-day SAN islands are typically fewer than one hundred ports, although many IT stores count on growth into a whole bunch of ports. SAN consolidation refers to both connecting distinctive SANs collectively or collapsing dissimilar islands into larger, more suitable-coordinated SANs with shared supplies.
additional stimulating the drive toward SAN consolidation are new disk array architectures that permit assorted purposes to take capabilities of varied entrance-end ports. this is one more reason businesses need to consolidate their information centers, servers, and storage. In doing so, they should still trust re-architecting their storage networks for the longer term. here is very true for corporations which are given that imposing "storage pooling" or "virtualization," which may complicate one of the most issues presently current in SAN deployments.
by treating SANs extra like LANs or WANs, storage network architects can function with a stronger interior profile and be able to discuss with company gadgets, management, and other IT instruments. this can aid them do a far better job of planning for SAN scalability, manageability, protection, and availability.
several change carriers are addressing some of those issues with a new category of "intelligent" switches, characterised by way of more desirable management, enhanced community segmentation, high availability, and scalability. however whatever extra SAN gear you believe, feel in terms of building a storage backbone community and deploying elements in a more controlled manner.
As businesses develop, so should their SANs. except lately, SANs have been now not very flexible, and most have been proprietary. but the trade is maturing and companies are introducing extra-bendy SAN products.
they are additionally working together in quite a few associations, such as the Storage network industry affiliation (SNIA), to develop standards. These are all first rate signs, but for SANs to consolidate and grow they need to be designed in a unique manner than they have got been to this point.
organizations need to keep in mind their application requirements after which discover a spine network with a tiered architecture. via treating SANs greater like LANs and WANs, they will no longer handiest produce more long lasting and expandable architectures, but also enlist more help and input from different IT groups and company devices.
Robert Strechay is director of technical marketing and chief customer recommend at Sandial systems (www.sandi alsystems.com) in Portsmouth, NH.make the most of your HBAs
by Mike Smith
to a few conclusion clients, a host bus adapter (HBA) gives most effective a standard actual and information connection from a server to a storage network. but storage administrators be aware of that HBAs collect information concerning the fame and actions of the entire storage area nework (SAN), and that tapping that suggestions can help them manage more storage, extra reliably, at much less cost.
A Fibre Channel HBA is like a computer on an add-in card, with its own processor and real-time working equipment. not like a network card, which with ease passes packets between the community and server, an HBA performs tons of the protocol processing work that a community card passes on to the server CPU and maintains track of the state of each I/O transaction.
as an instance, when an application on the server concerns a request to the storage community, the server operating system passes that request to the HBA driver, which applications the request, puts it in an I/O queue, after which transmits it to a storage equipment or change.
all over this method, the HBA driver maintains an awareness of the fitness of all SAN add-ons and links, which can play a job in reducing prices while increasing the reliability and efficiency of networked storage.
The up-front price of a SAN can be dwarfed with the aid of the ongoing fees of managing, troubleshooting, and upgrading the network. These prices rise each year as the quantity of information increases and the storage network turns into more complex as users centralize statistics. Yet with all these demands, companies are below drive to reduce their storage management budgets.
right here's how selecting the best HBA can help:
Simplified firmware updates—Updating firmware for a single HBA is not a complicated process, but making certain the firmware for each HBA in a SAN is at the right revision level may also be a frightening assignment—specially in SANs with lots of of servers. Centralized administration strategies can monitor firmware revisions on each local and far flung HBAs and may allow far flung and simultaneous firmware updates on as many HBAs as crucial. by means of eliminating the should improve firmware server-through-server and permitting directors to use a single tool to manage gadgets across the SAN, agencies can cut administration expenses and boost SAN uptime.
Driver compatibility across product generations—Compatibility of drivers across distinctive generations of HBAs makes it possible for storage administrators to set up the drivers essential to keep the SAN operating (or enhance its capabilities) devoid of going to the cost and effort of changing HBAs. It also allows for a gradual introduction of subsequent-technology adapters with minimal disruption. This simplifies management, protects hardware investments, and reduces the downtime essential to exchange HBAs.
decreased complexity—No storage administrator wants to scan diverse consoles to be trained what's wrong in a storage network, tons less bodily tune down failed contraptions. applications that supply administration of multiple types of SAN hardware permit administrators to cut back the variety of purposes required to control the SAN, which reduces complexity and administration time. as an instance, an administrator might use the same management utility to rezone a SAN and update HBA firmware. The merits are elevated administrator productivity and SAN availability and efficiency.
Centralized storage management—clever HBA firmware can enable managers to reset or re-initialize ports on all of the HBAs in a SAN, test records paths from a single element, and remotely improve firmware. This reduces administration costs and deliberate downtime.
issue detection/prevention—The precise-time monitoring completed through HBAs allows them to immediately notice actual failures of SAN links. it may possibly take a server operating gadget tens of seconds to respect the sort of failure. but the HBA can realize an issue inside a 2d and immediately notify application agents running on the server to redirect the request through a functioning hyperlink.
This reduces downtime by using dashing the fail-over procedure from a failed part to a working counterpart.requirements
Most HBA vendors aid the business-usual ordinary HBA API, which gives information akin to mannequin numbers, firmware revisions, and topology for HBAs, however not different SAN add-ons. despite the fact, HBA providers are partnering with storage subsystem, swap, and management application companies to make use of the assistance gathered through HBAs to give SAN-huge monitoring and administration.
Most HBA companies additionally aid management requisites such as the textile machine management Interface and protection necessities such as the Fibre Channel Authentication Protocol (FCAP) and Authenticated CT (normal Transport).
The greater power you might be below to carry cost-efficient, tremendously attainable storage, the extra consideration be sure to pay to your HBAs. via making full use of the information that HBAs collect about SANs, directors can reduce prices and enhance reliability.
Mike Smith is govt vice chairman of worldwide advertising at Emulex (www.emulex.com) in Costa Mesa, CA.