Network management

AdRem’s international approach to network monitoring

AdRem Sales and Marketing Director Jarek Jerzakowski gives a tour of the company’s brand new network management and monitoring product, NetCrunch 5, at Interop Las Vegas 2008. The product, sold around the world by AdRem and its resellers, automatically draws routing maps and physical maps of the network and allows you to create geographical custom maps to graphically monitor and manage your network.

PacketTrap shares vision of integrated network management, monitoring tools

At Interop, I sat down with PacketTrap CEO Steve Goodman, who showed me PacketTrap’s network management console and talked about the company’s philosophy of integrating disparate tools.

Perfect storm on its way for networking pros?

I flew back from Interop this morning, and my head is swimming from all there was to see and try to comprehend. One thing, however, is crystal clear: Today, networking is about a lot more than the network. When the hot topics at networking’s biggest conference are virtualization, application performance and green computing, it’s probably wise to start looking into new realms of technology, even if your current position doesn’t require it. Thing are bound to change quickly, and you may be in for more than you bargained for.

Dr. Jim Metzler warned attendees during the “Managing and Controlling Application Performance” session on Tuesday that in the next five years the combination of application acceleration, virtualization, and service oriented architecture will create a perfect storm. He said the resulting complexity of technologies may be something we are unable to manage. That’s quite an unsettling thought, especially coming from Jim, who may be the most influential expert in network and applications management today.

Taking a brighter view, other experts and vendors referred to the network and its new role as the “application delivery network” and emphasized the need for IT professionals that understand all aspects of delivering applications from end to end. These new application delivery specialists are most likely to come from a networking background, they agreed, because they have the knowledge of the complex infrastructure already in place.

So is the glass half empty or half full? Or maybe we just need to build an ark to weather the perfect storm. As Burton Group analyst Eric Siegel quipped to me at the end of the Metzler session, “You know what my plan is? In five years, I’m just going to retire!” If only we could all do that…

• Be sure to check out more Interop news and insight on our special coverage page
• Shout out to network management compadre Denise Dubie at Network World for her work with the aforementioned Eric Siegel on this useful guide to Assembling a ‘Network A-Team’

Open source networking hums along, quietly

Open source has been buzzing lately: Gartner identified it as one of the top 10 trends for 2008, and last week Microsoft announced it was publishing 30,000 pages of documentation for Windows Server 2008 and Windows Vista and launching an “Interoperability Initiative.” Google is pushing open applications and development of the new Android operating system, while Verizon claims to be opening its network.

In the networking industry, however, while open source adoption seems to be growing at a good clip, I don’t see much hoopla about it. So I’ll forgive you if you missed Shamus McGillicuddy’s article about the launch of ZipForge, a new website Alterpoint developed to support its ZipTie open source network configuration product. The ZipForge site provides a place where AlterPoint vendor partners can post interoperable ZipTie components that developers and users can download, review, and contribute to.

It would be great to see users take advantage of this repository to consolidate other networking-specific software tools, much like a true SourceForge (from which the new site partly takes its name) for networking pros. According to the article, networking has experienced less of an upsurge in open source because the technology itself is so reliant on hardware. While that may be true in a basic sense, software is becoming far more important and familiar. There are already several open source programs that networking folks use regularly, and that list is bound to expand.

Network engineers have long used open source software to run routers, VPNs and VLANs on run-of-the mill servers. Snort, OpenNMS, Nagios and Nessus are staples in networks big and small. And the popularity of Asterisk, the open source IP telephony platform, continues to grow in leaps and bounds.

Open source is definitely a part of the network, but I think that’s how most networking pros view it — as just a part of the network. They choose it because it works well, it interoperates, or it’s cheap, and they don’t get too caught up in the idealism and spreading the word about the benefits of open source. Also, most networking pros wear so many hats that they can’t spend a lot of time thinking about one system or product. They are even less likely to use that time evangelizing or flaming posts over at Slashdot.

I recall a network administrator I met a while back at a trade show. He had installed a few Vyatta routers, and he thought they were fantastic. But he was also in the throes of rolling out Avaya IP telephony to multiple locations and installing a new supply chain automation system. So while he was happy with his open source routers, they probably weren’t the first thing on his mind. The buzz about open source in the network is there, but sometimes you have to listen hard to hear it.

Network security concerns: Mo’ remote workers mo’ problems

When the network was built like a castle, located in one static location, it was easier to have perimeter defense–the castle walls, the moat and hill (i.e., the firewalls)–protecting the royalty… I mean, data. Nowadays, there’s more royalty (information) to keep track of and they don’t stay put within the safe walls of their core network/abode.

Needless to say, security is a primary networking concern (as was seen in a SearchNetworking.com survey conducted last fall polling more than 1,200 readers). I blame this largely on the increase of wireless (many wireless network security best practices are mysteries to most), the growing deployment of mobile devices (anywhere access), and the fact that not only are corporate devices travelling well beyond office-building walls, but the workers are too.

Sixty percent of enterprises have wide-spread remote access–where 50% or more of the workforce have remote access to the internal network–according to Yankee Group who surveyed 200 enterprises last summer. Senior Analyst of Enterprise Research in Network Security Phil Hochmuth of Yankee Group reported that three years prior, less than 25% of organizations supported wide-spread remote access.

With such an increase in such short amount of time, it’s no wonder network administrators are worried about how to secure and manage all these people. I realize I’m one of them: in the past three years, the companies I have worked for either allowed me to work remotely or involved me working from home entirely; I’m living this statistic, as many of you now are too, I’m sure.

Office space can be costly for an enterprise, and for the workers–so can gas, auto-repairs, and overall transportation. But on top of avoiding commutes, there’s a business benefit; mobile workers in jobs like sales or consulting, which require them to travel, are able to access data, fill orders more quickly, and quicken the overall pace of business transactions because they no longer lose as much time when they’re on the road. Unfortunately for the system administrator, all of this remote interaction puts stress on the network.

Hochmuth said “increased employee productivity is the main driver behind the move to open up internal networks for anywhere access, and SSL VPNs are emerging as the main tool enterprises use to provide this type of access.”

Independent research firm Amplitude Research commissioned by VanDyke Software found in their Fourth Annual Enterprise Security Survey that organizations are heightening their commitment to securing data communications. Secure remote access was the number one security management issue facing their company, according to their 2007 survey.

“The survey findings correlate to what we see happening in the field,” said Jeff P. Van Dyke, president and founder of VanDyke Software: “There’s a lot on the plates of the systems administrators, and with securing remote access a top issue and secure file transfer showing significant increase as a top issue to manage within the enterprise, VanDyke Software focused on new features for SecureCRT 6.0 and SecureFX 6.0 that make life in these areas so much easier for IT and network administrators.”

Hochmuth said “enterprises are literally opening up for business when it comes to supporting the ability of their employees to work from anywhere.” And VanDyke Software is one such company “opening up for business” to meet the needs of floundering network administrator’s who have to implement and support the increasing number of remote workers for their companies.

I suspect many more are aiming to follow suit.

Who is the “new” networking pro?

For more than a year, I’ve been hearing a lot of buzz about the concept of the “new” networking pro from networking vendors and pundits alike. The idea is, in a nutshell, that the old-school network guy spent his time thinking about network plumbing — connectivity, throughput, and Layer 3 hardware. Our publisher likens this guy to the Comic Book Guy from the Simpsons. (Don’t you work with at least one IT guy who looks a bit like that, after all?)

At the other end of the spectrum is the “new” networking pro, who is supposed to look like the guy below, in the suit. This person takes into account evolving technologies and network convergence. His concerns go beyond just plumbing; he concentrates on maximizing bandwidth, optimizing application and WAN performance; he troubleshoots a network that must accommodate voice and video traffic. The new networking pro realizes that he isn’t operating in a silo, but must cooperate (and, increasingly, share responsibilities and tools) with the security, voice, and data center staff.

Last week, two of the major network vendors — Cisco and Juniper — launched new switch offerings that are very much attuned to the “new” network pro.

Cisco’s Nexus 7000 super-switch may sound like it’s all about increased throughput, but perhaps the greater benefit, one which appeals to our new network pro, is that it is designed to eliminate redundant connectivity by allowing for converged Fibre Channel and Ethernet networks. And Cisco’s strategy, according to Doug Gourlay, senior director of marketing for Data Center Solutions at Cisco, is all about new network trends like convergence and virtualization (as quoted from an article by Shamus McGillicuddy).

Gourlay said Cisco has recognized several drivers that are leading higher-density data centers to turn to the network for help in improving operations. First is a “new class” of applications, such as voice and video, which consume significantly more bandwidth. The second driver is server virtualization. When you consolidate 10 virtual servers onto one physical server, that single box now deals with the aggregate bandwidth demands of the 10 servers. The third driver is Input-Output consolidation.

Similarly, Juniper’s new switches — the EX 3200, the EX 4200 and the EX 8200 — appeal to the network performance-minded enterprise, with their emphasis on uptime and reliability. As quoted in an article by Michael Morisy, Eddie Minkill, executive vice president of Juniper’s worldwide field operations, says that the view of the network as “plumbing” is outdated because consumers and businesses expect instant responses and more and more communications are carried through IP.

So the vendors and the pundits have their ducks in a row, but I wonder whether real-world network pros are getting in line. Surely these new switch offerings bode well for network reliability and productivity — but whenever I talk to actual people who are down in the IT trenches, I hear again and again that a) they would like to implement new technology, but instead spend more time keeping old stuff running and b) they don’t have time to think about anything besides putting out fires. (Maybe disgruntled IT pros are more verbal about this stuff.)

This skepticism isn’t unique. In the same article about Juniper’s new switch, Morisy writes that Jim Metzler, vice president of Sanibel, Fla.-based consultancy Ashton, Metzler & Associates, was doubtful about the current market for the powerful switch:

“I don’t know who they are going to appeal to,” he said, adding that most enterprises are not prioritizing carrier-grade robustness and millisecond latency. A few major exceptions exist, particularly in the financial sector, but he disagreed with the assertion that enterprises see the difference between 5 and 4 9’s of uptime as a critical tool in keeping customers.

“Is it good technology? Absolutely,” Metzler said. The problem is that people will not switch simply to use good technology, or even better technology, when what they have works, and Metzler said Juniper has found a solution for a problem people do not currently have. If the EX Series is to become a success, he said, a better marketing strategy must be implemented that could focus on cost reduction or other, more enterprise-focused problems.

Ultimately, I think that defining the network pro according to any stereotype is going to prove inaccurate, but the successful IT person may be the one who can look beyond their departmental silo and embrace “new” networking concepts like mobility and application performance. For that to happen, though, network pros may have to find a way to prevent those fires from happening in the first place.

Network management can be just plain ugly

I’ve been a writer and editor in the enterprise networking market for 11 years now, and one topic I dread being asked about is network management. I didn’t really understand it when I started learning about networking, and I don’t really understand it now. But the more I learn about the topic, the better I feel about that; because I meet few people who understand it much better than I do. 

The problem is not network management itself; the idea is straightforward and is of obvious importance, especially in companies where the network has a big influence on business success (which is most companies, these days). My challenge has been in attempting to understand and categorize a wide range of products that simply defy categorization.  

When I ask readers about network management, they often define it according to the particular tool they use, which can range from a basic protocol analyzer to a full-blown suite like HP OpenView. Within that range there are product groups that seem like they should be comparable, but features and functionality vary wildly. One network monitoring tool might show a schematic of your network and alert you if a device fails or traffic levels reach a predetermined unhealthy level. Another network monitoring tool is able to give detailed usage data for every device and performance metrics for each application running on the LAN and WAN, all presented in easy-to-read charts that compare actual results to past performance and service-level guarantees. 

This makes it difficult for network managers to choose appropriate products. Muddying the waters even more is the recent overlap between network and applications management, as well as security management and monitoring. Many products incorporate applications and security elements, making it even harder to compare them. Add in that enterprises typically use several network management products in a layered fashion, all connected to element managers for different devices, and you have a complex scenario.

How can you make sense of network management and use it to its fullest advantage? One way is to have a clear game plan. The report, Network management systems: The good, the bad, and the ugly, brought to you by our partnership with Info-Tech Research Group, provides a step-by-step methodology for evaluating your network management environment. The resulting analysis can help you identify where your existing products are already managing the network well, how they work together, and where you should be making investments in additional products or upgrades. Check out the report and let us know if it helps make your network management look any less ugly.

The problem of network management

Back in July, I wrote in a SearchNetworking newsletter about the challenge of choosing the right network management tool:

There’s not only overlap among these tools, but also a lot of variation in what they aim to manage — so much so that an apples-to-apples comparison is almost impossible to make. So far, it’s up to the people using the tools to clear up the confusion and determine the best approach.

I then asked readers to send me their thoughts, and received a very insightful reply from Internet security professional Rob Newby:

Network Monitoring and Management is a space which has been booming in recent years. A number of tools have grown up to monitor jflow, netflow, cflow, etc. There are more SIM, SIEM, and Log Management tools than I care to think of at present, most of them starting with “Net” something or “Log” something.

However, like all the simplest questions, “Why is the sky blue?”, “Why is there thunder and lightning?”, etc., the answer is longwinded and complex, and not as simple as this.

I have worked as an SE, and lately Product Manager for various companies, selling IT security tools, network add-ons, devices, software and hardware. For as long as I can remember, people have asked for centralized management and simple monitoring. The problem, of course, comes from the fact there there are no standards for these security devices and tools, apart from weak protocols such as SNMP and syslog, which are not up to the task of controlling and watching a network of hundreds of nodes.

To prove the lack of alternatives, HP Openview, an SNMP tool which gives a picture of network health by picking up SNMP traps, is still as popular today as it ever was. Nagios, and open source version is still used in many enterprise environments. Syslog collectors are available for all the “Net” and “Log” devices mentioned above.

Because of the lack of standardization, centralization is increasingly difficult unless you have some sort of vendor tie-in. Microsoft’s Operations Manager (MOM) is looking to be the most likely candidate for popular centralized management as the market crawls forwards to its decision. At the moment, it is wide open, however… and vendors are also moving towards SOA type offerings which can interact without the need for building APIs.

The market itself is crowded and becoming more complex. It is hard to make progress in any of these areas, and those leading the standards are the ones who will inevitably make the best of the markets. A common event format is being proposed for Log Management devices, for example. If you can’t standardize the software everyone is running, standardize the output — it makes sense.

Read Rob’s blog, “IT Security: The view from here” at http://robnewby.blogspot.com.