As companies migrate to New Data Center architectures, it stands to reason that they'll look at a wide range of vendors with new tools to help. Here are seven promising products.
Acopia Networks' Adaptive Resource Switch
As the volume of data grows, so do the challenges surrounding data management, from adding storage devices to increasing the staff to manage them. Acopia Networks' Adaptive Resource Switch (ARX) can help out. The in-band ARX front-ends network-attached storage (NAS) devices and other file servers, acting as a proxy for downstream clients. ARX provides a single global namespace for multiple file-storage systems, resulting in a virtual storage environment. Virtualization provides transparent data migration, load balancing and a tiered storage infrastructure that delivers on an information life-cycle management plan, says Tony Asaro, head of Enterprise Strategy Group's storage lab.
"NAS virtualization is a hot topic, and [large enterprises] all say they are either evaluating or considering Acopia ARX," Asaro says. Competitors are EMC, via its Rainfinity acquisition, and start-up NeoPath Networks, which is likewise getting traction, he says.
Acopia's customer list includes Boston.com, Goldman Sachs, Merrill Lynch, Toshiba and Warner Music Group.
Ibis Consulting, a Providence, R.I., company that helps customers with electronic discovery in response to litigation and regulatory compliance, has been using ARX products since January 2005. Ibis creates a virtual path name for each of its projects and assigns some amount of storage space, but it never knows for sure how much storage each project will require. Its two ARX switches have solved that problem, says Cliff Dutton, who had been Ibis CTO before joining Dynamic Communication, a management consultancy, early this year.
"If additional storage is required, it happens automatically," he says. "Ibis has avoided about 30% of the costs it would otherwise incur in managing the storage environment, which consists of more than 200TB of data.
Spreading data across multiple NAS systems eliminates I/O bottlenecks. And the ARX gear creates multiple copies of all the data processed for more back-up security, Dutton says, with no throughput overhead hit.
AmberPoint's AmberPoint 5.0
Enterprises that invest in a service-oriented architecture (SOA) often find their existing management tools aren't prepared to monitor such a dynamic environment. AmberPoint's SOA management software addresses a number of challenges, including performance analysis, exception management, validation of function and performance, and secure service delivery.
AmberPoint customers include Best Buy, Fujitsu, Kaiser Permanente, Motorola and Northern Trust. Many of the company's executives come from Forte Software, a vendor acquired by Sun that integrates enterprise applications.
"I like AmberPoint's approach," says Judith Hurwitz, president of the consultancy Hurwitz & Associates. "It's really done its homework to look at what it means to manage an environment where piece parts and components that you're applying to a problem change regularly," she says.
With competitors such as Actional and Blue Titan Software, AmberPoint must strike the right partnerships, such as with major systems management players and other leaders in the SOA space, including HP, IBM, Oracle and SAP, to maintain market leadership, Hurwitz says.
In the meantime, AmberPoint is helping MedicAlert identify bottlenecks and ensure its Microsoft-based Web services infrastructure meets service-level agreements, says David Harrington, CTO for the MedicAlert Foundation in Turlock, Calif. The nonprofit uses Web services to keep data in sync between clients' E-HealthKey USB storage devices and its central data repository.
"Implementing the service interfaces between E-HealthKey and our repository was really an exercise in connecting two legacy databases," Harrington says. "AmberPoint was the best dashboard of instrumentation for us to see how we were doing."
AmberPoint software also decrypts and encrypts messages in transit and provides a virtualization capability so that multiple MedicAlert Web services can be aggregated to look like a single service to outside partners, thus simplifying integration and adding security.
DataSynapse's GridServer
The idea of taking collections of low-cost computers and pressing them into doing the work of a mainframe fits well with today's do-more-with-less ethic, says Dan Kusnetzky, a vice president at IDC. With its GridServer Virtual Enterprise Edition, DataSynapse is focusing on one of the market segments where the technology is having the most success: financial services. Customers include Credit Suisse First Boston, Goldman Sachs and Wachovia.
"It doesn't require rocket scientists to make it work if you already know Java, VisualStudio, C++ or one of the other languages [DataSynapse] supports," Kusnetzky says. "It has tools to help people in [the financial services] arena develop and deploy applications."
DataSynapse isn't focused solely on financial-services companies; customers also include human-resources firm Hewitt Associates and Nationwide Mutual Insurance Co., for instance. But having some focus is important for a smaller company in the crowded grid market. About 30 vendors address various pieces of the puzzle, Kusnetzky says. Depending on the proposal, DataSynapse could run against HP, IBM, Microsoft or Sun, or a relative old-timer, such as Platform Computing.
Powerex, which markets wholesale energy products, has had impressive results with GridServer. The Vancouver company is using it to run a risk-analysis application from SunGard dubbed ZaiNet Monte Carlo Value at Risk. Jeff Gingera, IT director and security officer for Powerex, previously ran ZaiNet on an HP DL560 four-way computer. With ZaiNet now on the grid, an average report runs about 15 times faster, while the most complicated is about twice as fast. "One used to take six to 10 hours; it now finishes in 8 to 9 minutes," Gingera says.
DataSynapse faces the same challenge as any grid player: finding ways to segment older applications so that they can take advantage of grid computing, Kusnetzky says.
DataSynapse is well-positioned in that regard, Gingera says, because it operates at the suboperating system level. This means an application sees only one operating-system instance, not the multiple CPUs in the background. "That makes it easier to port applications to the grid," he says.
Enigmatec's Execution Management System
|
Enigmatec's Execution Management System (EMS) tries to minimize the number of people involved in managing networks. All sorts of management tools can send an alert when something goes wrong, but finding a fix often requires many decisions. EMS is intended to help organizations define the steps they take to address a problem - failing over an application to a back-up server, for example - then execute those steps without human intervention.
EMS works with its own network of agents and a company's existing management and provisioning tools to identify problems and deploy resources. It is particularly focused on tasks associated with NDC architectures, such as utility computing and virtualization. "It's a pretty powerful story," says John Humphreys, research manager for IDC, noting that EMS enables companies to better leverage management tool investments they've already made.
JPMorgan Chase, one of seven customers, uses EMS to address system failures quickly, says Shawn Findlan, vice president for global credit trading and global emerging markets for the New York firm. "If we have a failure in a primary data center, an alert will go off and the Enigmatec software captures that. In a few minutes, Enigmatec will automatically migrate that application to a new data center with adequate resources," he says.
Policy-based automation for SOA is a developing market, Humphreys says. Competitors include other young companies: Cassatt, Sychron and uXcomm. BMC Software, HP, IBM and Sun are expected to move into the market.
iConclude's iConclude Repair System
IConclude, too, is trying to address management problems in the New Data Center. Repair System, an agentless platform, automates the repair of some problems based on predefined scripts. And, it provides diagnostic aids to speed problem resolution by administrators or staff.
"This is a firefighter's-friend kind of play," says Dana Gardner, president and principal analyst at Interarbor Solutions. Rather than simply reacting to problems as they occur, Repair System "provides triage and automation."
The product also is intended to help customers define and adhere to repeatable problem-handling processes, such as those defined by the IT Infrastructure Library.
IConclude launched Repair System last November. Its only named customer is NSRI USA, a subsidiary of a large Japanese logistics company. The company is led by Sunny Gupta and other veterans of Mercury Interactive and its acquisition, Performant. "Sunny's had a good track record of success," Gardner notes. "IConclude can be given a serious look in an RFP process." Such a process could include a number of competitors, such as Indicative Software, Opnet Technologies and ProactiveNet, as well as HP, IBM, Microsoft and Red Hat.
NSRI began looking at Repair System because it is moving from two to a single operations-support center. "I'm looking for any way I can to automate functions," says Richard Dixon, vice president of NSRI.
The company came up with an immediate use for Repair System. NSRI often gets electronic data interchange (EDI) transactions that it can't match to its systems, usually because of an invalid customer or location code coming from an overseas region. Previously, a programmer would fix the data. Using Repair System, the company developed a script to analyze problem EDI transactions and determine whether they have a customer- or location-code problem. If so, Repair System generates an incident report in HP OpenView Service Desk and sends the source an e-mail describing the problem and how to fix it. "[IConclude] totally took away the need for a programmer to get involved," Dixon says.
Mirage Networks' Mirage NAC
Network-access control tools are often touted for their ability to ensure that potentially vulnerable clients in remote locations don't connect to a network. Mirage NAC takes that concept a step further.
It not only conducts preadmission checks, using the McAfee Foundstone Vulnerability Management System, but also continually checks for anomalous behavior while clients are connected. Any offenders are isolated. Mirage detects anomalous behavior by maintaining a map of unused IP addresses and sending an alert whenever a device tries to access one of them, says Chris Liebert, senior analyst with The Yankee Group. "It's a good approach," she says.
Customers include high-tech companies National Instruments and OnDemand Software, law firm Hogan & Hartson and the Pennsylvania State University. Users report Mirage NAC has dramatically decreased the time it takes to find problem devices on their networks, Liebert says.
Chris Hanson, IT project manager for Kern Schools Federal Credit Union in Bakersfield, Calif., is one such user. What he likes best about Mirage NAC is that it's agentless. "So many security products want you to have their special client to watch this, that and the other," he says. "Pretty soon it becomes a nightmare."
Competitors include Arbor Networks, Cisco, Lancope and Mazu Networks. Mirage's biggest challenge is to improve the NAC's reporting capabilities and to make it more selective about what alerts it reports on, Liebert says.
Softricity's Softricity Desktop
Managing Windows applications is no picnic, between the servers that run various components and the potentially massive amount of often-flaky client code on each desktop. The idea behind Softricity Desktop is to get rid of most of that code by centralizing all applications and having them delivered to clients as services on an as-needed basis. If that sounds like the old thin-client song that Sun CEO Scott McNealy has been singing for years, it pretty much is. The key difference is that Softricity makes it work for Windows applications, such as Office and Exchange, that most companies rely on day to day.
"It's really about managing the complexity of Windows," says Interarbor's Gardner. Softricity Desktop eases updates, patches and deployment when managing dozens to hundreds of Windows applications, he says.
That's been the case for Heartland Financial USA, a bank-holding company in Dubuque, Iowa, that as of late last year was supporting about 700 of its 1,000 users from the Softricity platform. The help desk now spends about 80% less time grappling with application issues, says Marti Vandemore, vice president of IS. "Now, we know the application works," he says. Even better, he adds, "in just a few minutes, we can deploy [an application] to 100 users."
The product also simplifies disaster-recovery planning, because server images are independent of the hardware on which they run and can be restored easily on another server. "The time and resources that we're saving have far outweighed the cost" of the Softricity software, he says.
Among the challenges Softricity faces is that it works only with Windows applications, Gardner says. Another drawback is that one virtualized application can't call another, such as when an Office application wants to invoke an Adobe Acrobat reader, says Shane Nicely, assistant vice president of IS at Heartland Financial. His overall impression, however, is that Softricity has "worked much better than we expected."
Desmond is president of PDEdit, in Southborough, Mass. He can be reached at paul@pdedit.com. Senior Editor Denise Dubie contributed to this story.
Previous: Credit to the New Data Center | Next: The New Data Center desktop