Network Management For The Modern Network

One company making strides in this direction is AccelOps, a three-year-old Silicon Valley startup founded by former Cisco Systems engineers. AccelOps’ flagship product, and namesake, is designed to gather and interpret data from device and server logs, track configuration changes and display the differences, visualize device and application dependencies to simplify problem diagnosis, and reduce downtime.

AccelOps became generally available in June 2009 and competes with solutions from ManageEngine, Nimsoft (which is soon to be acquired by CA Technologies), and SolarWinds. AccelOps unveiled a channel program this month that permits solution providers to package the AccelOps service as their own.

AccelOps is aimed at small and midsize businesses with 500 to 5,000 users, a market that, according to Scott Gordon, vice president of marketing and business development at AccelOps, lacks a complete solution.

“Most of their process is around configuration management and help desk. Everything else is piecemeal. So system integrators and VARs have an opportunity to offer managed services,” said Gordon.

At the solution provider’s discretion, AccelOps typically sells for around $2,000 per month, of which the solution provider gets a cut, said Gordon. The multitenant Software-as-a-Service Modern Network for today’s businesses, it’s all about intelligence edition requires an on-site dual-core server running VMware’s ESX or ESXi. Client-side software performs local discovery and collects data, which it compresses, encrypts and sends to the AccelOps back end for processing and storage. An Internet connection is required.

Costs, which are the same for the SaaS and VM versions, are based on the number of configuration items, which include operating systems, applications and anything with an IP address, as well as the number of events per second (EPS) they generate. In contrast, competitor SolarWinds charges by the process which, according to Gordon, ends up being more costly.

“Every time you launch an application and it kicks off a series of processes, there’s a charge. With AccelOps, we only charge the first time we see the application. After that, it only counts as one [configuration item].” Also unlike competitive solutions, AccelOps clearly displays the number of configuration items being counted. “And we let you pick and choose what [configuration items] to keep. And when you relicense, we pro-rate the difference.”

When deployed as an appliance, AccelOps requires a four-core server with 8 GB of memory and 300 GB of available hard disk space. When two AccelOps licenses are available, performance is improved automatically.

“When you buy two licenses, they see each other and automatically load-balance,” Gordon said. Competitive solutions, he added, require such capability to be configured manually.

Modules in the extensible AccelOps solution include network behavior analysis; a configuration management database with Layer 2/3 topology mapping; agentless discovery; resource and process analysis on all connected systems; extensible Boolean, nested and statistical profiling and roles-based access control. Data is encrypted.

The CRN Test Center was unable to install and configure the self-hosted edition of AccelOps in time for this review. Instead, reviewers were given access to an AccelOps site and worked with the software remotely, accessing it from a variety of browsers on Mac OS X and Windows. All worked flawlessly. What we also found was an extremely powerful, stable and comprehensive monitoring and management tool that would benefit nearly any organization, particularly those with numerous network assets that are varied and widely dispersed.

We’ve only got space here to describe a few of the most useful capabilities. The first is for that “the application is slow” problem. Most topology maps do a fine job of finding and displaying devices. Heck, even Windows can do that. AccelOps stands out for the way it displays the devices involved in a particular service, along with their status. Select the dashboard for a service and all devices involved in that service are connected by red lines (see screen shot) along with the number of device warnings, if any (also red). Drilling into a device displays details about recent incidents, device health, and stats on relevant interfaces, services and apps.

Providing higher-level views are summary dashboards for apps, availability, performance, incidents, security, devices and business services. Selecting a service from the latter, for example, displays lists of monitored devices and apps in the service along with statistics on uptime, performance memory and processor utilization, as well as a variety of other system data. Red items stand out as critical alerts, yellow items signify warnings and green items signify go.

NEXT: Help For Administrators

Help For Administrators

Gordon points out that AccelOps is not a help desk, but does include a task manager that can be “used to create tasks for events within AccelOps, and there are some who use it that way.” Administrators will appreciate its password management capability, as it is able to automatically update its device password database as compliance rules might dictate, for example.

“If you’re compliant, you’re changing passwords every 90 days. With AccelOps, you don’t have to go in and change them on each and every device. You do it once and it’s done automatically,” he said.

As with any powerful centralized management tool, setup can be a bit hairy. Included is a preconfiguration service whereby AccelOps inputs all known and supported devices prior to deployment. The tool discovers the remaining devices, reads and stores log and other device data and can still provide useful information.

“If you send us your data, you don’t have to set anything up; we’ll take the data and start parsing it,” said Benjamin Powell, director of services at AccelOps. “If people don’t let us discover a device, we can still learn from it and understand some things that are happening,” said Powell. “We might not know about memory and disk status, or what users are doing and which interfaces they’re communicating with. But if you allow us to communicate with the device and get its logs, we can still give you a picture of what’s happening,” he said.

Among AccelOps’ most useful capabilities is its ability to present a complete picture of what’s happening in any compute scenario across all company boundaries. “I can’t get at my Web app,” a distressed user might say. Beyond pinging the server, AccelOps lets administrators generate a synthetic transaction and watch reactions of all affected systems from end to end in realtime.

“If you can see all the pieces that are involved, it’s easy to see which is the squeaky wheel,” said Powell.

If something has changed, differences can be displayed for an easy fix. For network administrators, help desk and customer service staff, and solution providers performing such services on behalf of customers, the CRN Test Center highly recommends AccelOps.

NEXT: Managed WAN Managed WAN

If you’re looking for an opportunity to optimize and manage data transport services (and increase recurring revenue), Riverbed Technology might be worth a look. As data centers grow, consolidate, virtualize and become ever more geographically dispersed, maximizing WAN throughput can mean life or death for a company. And such companies will usually pay handsomely to minimize risk of network failure.

Riverbed’s flagship product is Steelhead, a network- and application- layer appliance that when installed at both ends of a WAN connection produces bandwidth performance improvements of at least five times and as much as 100 times, according to the company. Although the Test Center could not put those claims to the test, we were able to provide some insight into the way the technology works.

Running inside the Steelhead appliance is the Riverbed Optimization System (RiOS), the company’s proprietary secret-sauce operating system for minimizing network traffic. At the heart of RiOS is a universal data store that tracks and archives binary data as it passes along the wire. After that, repetitive data is replaced with a reference.

“For example, your company logo might cross the network many hundreds of times in a day. We see it once and just send a reference to it after that,” explained Joe Ghory, Riverbed’s senior product marketing manager. Unlike competitive solutions from Cisco and others, which analyze data at the block level, Riverbed’s binary analysis is protocol-neutral.

“We work at the byte level -- reading ones and zeros, not as block-level data, which is less effective,” he said. Also unlike competitors, which employ peer- or user-based data stores, “We don’t discriminate as to where the data is coming from, just as to whether we’ve seen it before. And by having a universal data store, we optimize for an entire organization, regardless of the protocol.”

In June Riverbed began shipping version 6.1 of the Riverbed Optimization System, a free update for maintenance customers that added optimization for Lotus Notes 8.5, Microsoft Exchange 2010 and SharePoint, as well as their corresponding online versions. For the department or branch office, there’s the optional Riverbed Services Platform, which permits as many as five dedicated VMware virtual machines to run within a Steelhead appliance, delivering file, print, DHCP and whatever services are needed there.

WAN optimization provisioning is not for every reseller. To qualify as a service provider partner, Riverbed requires the prospect to operate a NOC and engage in the design, construction, implementation and management of network services. Existing partners include AT&T, Orange Business Services, Verizon and other large telcos. No such restrictions exist for system integrator partners, which the company supports with an aggressive lead generation program as well as training on sales, Level 1 and Level 2 support, vertical-market specialization and professional service provisioning.

End-user pricing ranges from $6,000 to $250,000 per device and is dependent on the number of TCP connections, the size and number of files to be optimized and the desired throughput.

NEXT: For Small Business, New Business For Small Business, New Business

The idea behind integrated network management is straightforward enough so that anyone in any enterprise should be able to get it. Keeping 24x7 tabs on all devices on a network, all potential weak links and threats, all potential drains on performance, are essential best practices for even small networks.

The problem with advanced network management, like all IT, is in convincing those who own or run smaller networks that they need every bit of the protection and network optics as large networks -- especially given the ever-tightening compliance regimes that emerge.

Enter Spiceworks, a privately held software startup based in Austin, Texas, that has provided some of the sophistication of traditional network management solutions with an important, added benefit for many: It’s free.

We installed Spiceworks on a PC built with an AMD Phenom 9550 quad-core processor at 2.20GHz, running Windows 7 Ultimate 32-bit. The dashboard and management interface are both Web-based -- and ran fine using Firefox 3.6.

Within two minutes the software, via SNMP, audited the network and performed an asset inventory of all connected devices -- including workstations, PCs and peripherals. The included Network Map is designed graphically to be just that: a map of everything that’s connected to the Internet. Clicking on each device on the map brings up information on each device including model, serial number, IP address, MAC address, netmask, DHCP, DNS, domain, OS, BIOS and location.

What really impressed us was that Spiceworks is VM-aware management software. Not only did it include virtual machines in its asset inventory, it even told us of an error with a VMware ESX server. (The error turned out to be an expired license.)

The console also provides device monitoring. While less robust than many paid-commercial solutions, Spiceworks monitoring allows an administrator to set alerts for any number of preconfigured events, including installation of software like Weatherbug, Google Desktop or other software that administrators may believe could pose a threat to stability or compliance requirements. It also allows for a menu of customized e-mail alerts for software events, hardware events, printer supply issues, hotfixes and more.

Think of Spiceworks as a great gateway for small or new businesses into the world of best practices and network management. The more network data at the hands of these businesses, the more business owners are likely to want. The CRN Test Center can recommend Spiceworks for new and small businesses -- and we look forward to seeing this software grow and improve over time.

COMMUNITY: Connect with the CRN Test Center at community.crn.com.

Startup company AccelOps' flagship product, and namesake, is designed to gather and interpret data from device and server logs, track configuration changes and display the differences, visualize device and application dependencies to simplify problem diagnosis, and reduce downtime.