Enterprise Network Management

Network Management is a big term that ultimately encompasses all factors related to a network, be it an enterprise or a service provider network: Security, monitoring, planning, topology mapping, analysis of network performance to safeguard against traffic congestion of real-time communications applications — the list goes on and on.

As IP Communications evolves into new services and continues to infiltrate the enterprise, network management, out of necessity, evolves along with it, becoming more sophisticated and “intelligent” in the process. Take eTelemetry (http://www.etelemetry.com), (news - alert) for example, which develops products that, in a sense, passively turn network packets into business intelligence. Their products provide administrators with a ‘People Layer’, or the exact name and location behind IP addresses. IP phones can also be locked to a specific switch port and the system can automatically request verification from the phone’s assigned user, who must confirm a move before a phone will function at its new location. The company’s Metron product tracks bandwidth usage back to each employee on the network, ferreting out Web surfers and chat sessions during working hours. Metron also detects secure sessions initiated by employees that could potentially be used to transmit proprietary company data.

Similarly, Netcordia’s (http://www.netcordia. com) (news - alert) NetMRI is a next-gen network management tool designed for contemporary large, complex, and heterogeneous networks. Netcordia’s VoIP analysis module, together with the NetMRI appliance, serves as a comprehensive, automated network diagnostic tool in the industry, analyzing wireless, VPN, VLAN, IP, and VoIP problems. NetMRI gathers VoIP Call Detail Records (CDRs) from PBXs by such well known companies as Cisco and Avaya, analyzes the records on a daily (or even more frequent) basis, and generates a Path Diagnostic Chart. Such a combination of call information, call performance, and network analysis can be used to help troubleshoot specific VoIP problems. NetMRI also uses SNMP to capture data devices along and across network call paths, evaluating the device QoS configurations. If a given Voice Path doesn’t have QoS configured, or if the configurations change, an Issue is generated to alert the engineer or technician.

Over at Qovia (http://www.qovia.com) (news - alert) a provider of IP telephony monitoring and management solutions, Lou Nardo, Director of Product Management, says, “Qovia’s IP Telephony Manager is a software solution built from the ground up to monitor real-time IP applications (predominantly VoIP, (define - news - alert) but expanding). The system provides realtime monitoring on the status and health of all IP calls throughout the network and correlates the results with other environmental parameters in the voice ecosystem, such as signaling information, call server statistics, as well as application, server, and network device performance and health. Qovia differentiates in three ways: First, unprecedented real-time visibility into all calls down to individual hand sets; second, rich analysis and correlation for meaningful alarming, root cause analysis, and capacity planning; and third, rich and extensible reporting.”

“The Enterprise Network Management field is at a point where it needs to adapt to changes in the way IP networks are used,” says Nardo. “The rapid emergence of real-time applications as mission critical enterprise systems is applying enormous pressure on IT Operations organizations to keep up. Systems management vendors will need to find new approaches and solutions.”

One might wonder if all of these new approaches and solutions imply that managing a packet network becomes more difficult than a circuit-switched one. “I wouldn’t say more difficult,” answers Norda, “but certainly different, with its own set of issues to address. Adding to the problem, though, is the newness of the challenge to packet networks. IP telephony is a real-time application, and, although packet networks have been managed successfully for quite some time, new management techniques are now required. Unfortunately, many network administrators assume they can continue to apply their existing tools and troubleshooting techniques and successfully manage an IP telephony deployment. Many find out, only after significant deployment problems, that they need help. Since packet networks chop up the call streams and send them packet by packet, the system needs to be able to accurately reassemble the call. All of a sudden, minor packet loss and delay issues that were easily tolerated by traditional IP applications have become significant challenges. Network administration teams are faced with changing their views on network performance and their approach to management.”

“Maintaining quality of service could be more difficult than other network management activities,” says Norda, “but that’s why we put so much effort into providing an easy-to-use system with plenty of built-in expertise. Maintaining quality of service is challenging for many reasons, not the least of which is how subjective it can be. Perception is reality and often unfair, but you need to manage it nonetheless. The trick is to get in front of it. Be proactive and eliminate emerging issues before they ever impact the user experience. That takes some significantly granular, real-time visibility into the health of the systems, but also the leading indicators to problems. Monitoring these leading indicators and intelligently correlating them with their impact to quality empowers you to see and address problems before your customers or end users are impacted.”

Allot Communications (news - alert) (http://www.allot.com) provides intelligent IP service optimization solutions based on deep packet inspection (DPI) technology. Allot’s NetEnforcer devices monitor and shape broadband traffic; Allot NetXplorer Management centrally collects and reports usage statistics and defines control policies, while the Allot Subscriber Management Platform provides subscriber awareness needed to customize service offerings and maximize service revenues.

Allot’s President and CEO, Rami Hadar, says, “Allot is a sort of dual company in that our products are relevant both for the high-end enterprise market and service providers, such as ISPs, DSL and cable operators, and mobile operators. We employ advanced deep packet inspection or Layer 7 technology. We sit in tandem to IP links, whether it’s one link outbound from an enterprise or an uplink from an aggregation device. If it’s a DSLAM in a DSL network or a CMTS in a cable installation, we sit in tandem in front of edge routers. We’re unique in that we inspect IP traffic in real time using deep packet inspection — beyond Layer 4, which deals with addresses and ports — analyze the traffic, and identify applications. We are able to look into a packet or series of packets and identify the applications that are carried. ‘Real time’ for enterprises ranges from hundreds of megabits per second up to one gigabit. For service providers, of course, we handle greater bandwidths, up to five Gbps.”

“Our proposition is based on three elements,” says Hada. “First, there’s knowledge resulting from monitoring. Many organizations, both enterprises and service providers, are not aware of what their pipes are carrying at any particular moment. We break down the pipe into applications and show the enterprise the distribution of how much VoIP traffic is there, how much peer-to-peer traffic is there, how much SIP traffic is there, how much gaming traffic goes on, how much downloading, and so forth.”

“Additionally, our NetXplorer product does provisioning and reporting,” says Hada. “It provides reports, both realtime and non-real-time, and long-term or short-term. It gives you information on the distribution of applications, statistics, and so forth. With our product you can see traffic jams, drill down to the user level, and see such things as who’s using the network and how, and which applications are causing congestion on the network. This type of monitoring is very useful to plan networks or extensions to networks. It also can be used to mitigate security attacks as they happen. Denial of Service [DoS] attacks can be characterized by many connections that someone opens in the network, and we can identify that automatically and react to it. We don’t need to specifically analyze what kind of DoS attack is being launched, but one can see very quickly if an unreasonable amount of traffic is occurring, or an unreasonable amount of connections are taking place from a certain user.”

“Once our customers see how the network is being used, they want to optimize the network, to be able to control how much network resources or traffic or bandwidth is apportioned to one application over another, says Hada. “For example, if your enterprise has just bought a VoIP PBX, you might want to give the VoIP traffic higher or guaranteed priority over any other packet traffic on the network, such as transactions, end-of-the-day backups, or employees listening to streaming music over the Internet. Not only can we identify such applications but we also enable our customers to now control just how much bandwidth is given to whatever application is considered important.”

The Trend at Trendium
Trendium (news - alert) (http://www.trendium.com) is a global provider of service assurance and performance management software deployed in both Fortune 500 and service providers. Trendium’s extensive portfolio of solutions includes its ServicePATH™ Service Intelligence System (manages services and service level agreements, or SLAs) and PerforMAX Performance Management System. The company’s new Service Bureau is a subscription-based “software as a service” line of business.

Tom Leh, VP of Product Development at Trendium, says, “Trendium focuses on Tier 1 service providers and significantly sized enterprise environments. We have two flagship products that have been in production for the past five years. ServicePATH is a framework software platform that can be abstracted into a number of different solutions and is focused more on the Tier 1 operators and big service providers with complex back ends and hundreds of different types of network elements and OSS and BSS systems with which to integrate. Our PerforMAX package is more enterprise oriented and focused more on the performance management function. From the higher layer of what is typically called in the industry ‘service management’, we look at the service level agreements and the service level objectives, and actually correlate the performance with contracts for a specific customer. At the enterprise level, we deal more with key performance indicators and we correlate those with different systems already in place.”

Looking now more at the enterprise, more recently, we have launched a new line of business, which we call the Service Bureau. Several different services will be offered under that name. The first has been dubbed “BroadSight,” which is essentially a service assurance package for DSL. It helps you to identify different types of hidden network anomalies to ensure that the customer experience is what’s expected so that an operator can reduce churn, particularly during the first 30 days, a time period known that’s used to calculate the sales to gain ratio.

Trendium’s Service Bureau will also include InfraSight™, a state of the art infrastructure health/performance management service and SLA Manager that will provide a portfolio of business related SLA/SLO service assurance for Ethernet, triple play, MPLS, WiMAX, and other managed services.

Leveraging Intelligence and Experience
Although we’re not yet close to developing fully automated, self-healing networks, network management functionality is improving in leaps and bounds, often in creative new ways.

At NextNine (http://www.NextNine.com), (news - alert) its Support Automation Software, leveraged by telecom leaders such as Motorola Connected Homes, Comverse, OpenWave, and others, by virtue of its automated, proactive, preventive nature, empowers organizations to drastically improve the level of customer service they provide, thereby increasing customer satisfaction and, by extension, customer retention, in addition to protecting service maintenance revenue and increasing profit margins. NextNine’s automated, proactive support solution enables organizations to shift from current, reactive support methods to a customer-centric, proactive support approach enabling early problem detection. Thus, network support professionals can now resolve problems at the root and save time usually spent on customer interaction sessions, lengthy data collection and complex diagnostics. The result: maximum network performance and uptime.

Gil Levonai, NextNine’s VP of Marketing, says, “We’re basically in the field of service for technology. It’s quite a wide field that has lots of verticals in it, such as VoIP and IP Communications, and encompasses both the enterprise and carrier sides. We work in healthcare, in other technology arenas, in industrial environments, and so forth.”

“Basically, our key component is what we call the Virtual Support Engineer,” says Levonai. “This is a piece of software that imitates what a human support engineer does. It’s downloaded on demand or permanently installed at the customer site. A single Virtual Support Engineer can automate the service, support, and maintenance of a single or multiple installed software and hardware systems.”

NextNine’s Virtual Support Engineer communicates with systems via various protocols and can access files, run commands and queries, scan logs, as well as actively fix, patch, and update applications, devices, and systems.

Levonai elaborates: “Instead of what you find in your ‘normal’ management systems — where they basically wait for events that come out of various equipment and process and correlate data and understand from the stream of data what’s going on — we take a different approach. We recognize the 20/80 rule; this is where 20 percent of the issues that can happen generate 80 percent of the actual incidents.”

“With companies such as Comverse (news - alert) and Motorola, (quote - news - alert) we’ve been able to help them reduce the amount of customer downtime by nearly 90 percent within a few months of deployment,” says Levonai. “That’s because we take the approach of not listening to the raw data of what the product is telling us in the sense of, ‘okay an R&D guy is creating SNMP traffic and now we’re listening to a stream of SNMP traffic and we’re trying to figure out what’s going on’. Instead, we’re actually asking the service engineers, when they go onsite and they try to analyze the problem, what kind problems and phenomena do they normally encounter and how would they go about avoiding or fixing them through actual understanding of what’s going on with the equipment and the network by identifying the signs.”

“Now, ‘identifying the signs’ can mean many things,” says Levonai. “It can be checking for connectivity in a couple of units; it could entail the known behavior of a router, a PBX, whatever. It can be performance checks on equipment, but all based on what the organization says has normally gone wrong. In effect, this approach uses ‘best practices’ of IT and service organizations and automating this expertise. It’s not just about listening to whatever the equipment is telling us. It’s an approach that works well when you have mission critical systems, as in data centers and high-end networks, where basically the behavior of the system as a system has known issues and you can find them proactively by looking for variations.”

And the Prognosis is. . .
Integrated Research’s Prognosis (http://www.prognosis.com) product range is a broad suite of monitoring and management software, designed to give an organization’s technical personnel insight into the health and performance of their key computer systems, and the business applications running on them.

Nathan Brumby, General Manager, IP Telephony and Voice over IP, says, “We’ve been very focused on the voice component of network — the true VoIP component — for the past six years. There are some interesting data points. The first are early adopters, the enterprises that adopted this about two years ago and are now truly starting to understand what they’ve adopted, the realities of it, and the set of problems that are different or distinct from what they had when they worked in a traditional telephony environment.”

“Secondly,” says Brumby, “we’ve seen on a global basis large managed service providers that have now reached the point where they’ve very clearly been able to articulate where they want to go in this space and how they want to sell value and services to their end users.”

“What the last two years have shown is that there are definite issues in migrating over to VoIP,” says Brumby. “It’s not just another application on a data network. It’s far more mission critical in many cases than, for instance, an email heading across a network. With VoIP, a learning process has taken place, centering on just how different it has been or is going to be. Companies, in general, felt that, when they migrated to VoIP, they were going to save a lot of money and they wouldn’t have to pay a big monthly check to a carrier. That cost on the infrastructure was supposedly going to be removed and they, probably quite naïvely, assumed everything would be fine. But, they now see how different it is to manage VoIP as opposed to the management of an ordinary data network; they now realize that they paid a lot of money to large carriers for a good reason — namely, if you want to have six nines of availability, if you always want to have dial tone, if you always want your message service to work, and if you want to make sure there’s no toll fraud, then all that requires a specific set of tools and a specialized set of capabilities that don’t immediately exist in a new market.”

“So, a maturing process has been going on,” says Brumby. “We’re reached a point now where these companies that were early adopters are now starting to realize that IP communications is complicated. Secondly, because it’s a young technology, there are certain things that the companies had in a traditional telephony environment that they don’t have with IP: least cost routing, automation of provisioning, and so forth. As a result, they’re starting to realize that they’re going to need more experience to be able to get back to the guaranty of quality that they had in the traditional telephony environment.”

“So that’s where we are today. There’s this realization about what companies truly have to do to have a good, robust telephony environment,” says Brumby. “Vendors individually often offer some set of management tools to help, but they might not be as robust as a third-party vendor, such as Prognosis. That’s because there are some conflicting messages that ultimately get put out to an end user. Take a vendor that wants to sell you an IP telephony kit. They’ll say to you that it’s robust, it’s reliable, it’s the best of the available vendors’ products in the market and you have to buy that product. Then they’ll conflict themselves a bit and say, ‘Well, look, we offer management tools and really, you should buy them’. But management tools are probably not their area of expertise. Therefore, you can obtain tools from vendors, but they’re generally not very comprehensive, because that’s not what the vendors’ business happens to be. Their business is actually just moving boxes, selling routers, and so on.”

“The situation is better when thirdparty products arrive from companies that have chosen to be experts in enterprise network management,” says Brumby, “and have chosen that as the point where they’re going to compete. Invariably, they do much better at it than companies such as ordinary vendors. Prognosis, for example, currently resides in seven of Cisco’s top ten deployments globally. On many occasions, in large deployments that are highly distributed and highly complex, Cisco will take us into those accounts, because, at the end of the day, their key concern is that their hardware, their ‘kit’, functions appropriately and they get paid for it and they get ongoing maintenance for it. Prognosis is part of that assurance.”

To sum up, in the past, enterprise network management was as good as your IT staff — sometimes great, sometimes bumbling in nature. As networks become more dynamic and complex, the ability to view of everything important occurring in a network at any given moment — identification of data bottlenecks, alarm information, outage identification, etc. — as well as the ability to model and plan networks for expansion or change, will become more important than ever, perhaps beyond the capabilities of mere mortals.

Fortunately, an army of network management hardware and software companies will be waiting to help.