Business Recovery For Data Networks
Solutions for Private Line & Frame Relay

By Randolph A. Fisher, CBCP


Data networking needs have been rapidly expanding as a result of new applications and higher bandwidth requirements.

Today’s data networks no longer use a single technology. Rather, they are hybrids using technologies that are being driven to the limit with more users and ever-increasing traffic.

At the same time, corporations have come to realize that their proprietary information is a most valuable asset, which provides a distinct competitive advantage. As such, it needs to be guarded, yet be readily available to authorized personnel. Industry regulations, court rulings and political unrest also remind us that data must be protected.

There are many methods that provide network redundancy and reliability. This article will briefly discuss some of the solutions offered by interchange carriers (IXCs) for private line and frame relay technologies.

Private Line Network Management

Sophisticated network monitoring and management tools have been available for many years. However, they are now gaining wider acceptance for several reasons:

1. Users have become more comfortable in taking responsibility and accountability for their networks.

2. Businesses—large and small—are now realizing that internetworking products can reduce downtime and thus could pay for themselves in six to nine months.

3. Network management tools give users on-premises control and monitoring capabilities.

Monitoring and Management Tools

What to Look for…
With these tools, users can monitor their network’s health, vendor performance and responsiveness, as well as optimize their services to meet dynamic time of day or date-sensitive application requirements. This is accomplished by partitioning their dedicated T1.5Mbps inter-office circuits into separate channels, thereby allowing multiple applications on a single facility. The result—more efficient circuit loading and better value!

Users can configure their physical circuits into logical channel groups and monitor and dynamically manage these channel groups from the DS0 through DS1 bandwidth levels. A further benefit is access to a shared network facility pool for additional part-time bandwidth.

Look for…
Numerous automatic restoration options that should be available for failed channel groups:

• Restoration on customer-owned spare bandwidth
Restoration to an IXC bandwidth- on-demand (B-o-D) pool
Selection of a specific route
Provision for audible and visual alarms with manual redirection

Look for…
Availability of these features throughout the IXC offices in the USA, Canada, Puerto Rico and Hawaii. The network management services should be supported by a dedicated service center, which is staffed 24 hours a day, 7 days a week. The center should have dedicated technicians trained to respond to disaster recovery situations.

Ask the question: "To what am I subscribing?"
Be certain that you are getting what you want. Potential subscribers should create a comparison scorecard based on attributes that will satisfy their specific application need and performance requirement. The minimal performance requirements are responsiveness, restoration speed, ubiquity, and ease of use. Price, although important, must not be your primary consideration for disaster recovery. The price of downtime far exceeds the cost for these tools.

Network Management – "Hand-off" or "Hands-on"?

Customer Managed Network
IXCs are offering network management tools for ‘hands-on’ or ‘hands-off’ preferences. Users with very dynamic applications needs usually request customer managed capability. Those with virtually static configurations usually request the vendor to perform functions on their behalf.

Customer managed networks utilize a PC or workstation with their IXC’s proprietary software to allow dynamic reconfiguration and control of permanent dedicated bandwidth. Connection to the IXC’s system controller and switching devices is with a dedicated 9.6Kbps or 56Kbps private line access channel. A dial-up access line is also available for those with virtually static networks that still want the comfort of knowing they could make their own changes.

The PC is used to communicate with the carrier’s system controller, which verifies security information. The system controller in turn communicates with Digital Access Connect Switching (DACS) devices deployed at the IXC’s central offices. These switches receive their cross connect information from the system controller and perform the requested network cross connects using either user owned bandwidth, or IXC owned bandwidth from the B-o-D pool to make the necessary connections.

Vendor Managed Network
The most widely used method for redirecting traffic from a primary location to a secondary location is the use of B-o-D by an IXC to reroute channels on the customer’s behalf. This is chiefly because customers:

Have a virtually static network
May not want the expense of 24 x 7 coverage and software training requirements

B-o-D is a shared pool of IXC facilities used for temporary peaking or disaster recovery needs. B-o-D is a very cost effective alternative to purchasing permanent dedicated facilities in many instances. Here are a few examples of B-o-D uses:

A large merchandise retailer sends receipts to its data center for processing once a month
A car manufacturer sends large design files to its assembly plants with the latest updates
A bank wants to run its semi-annual disaster recovery exercise

A Word of Caution…
Be mindful that there is a tradeoff – risk versus price! If you need guaranteed bandwidth availability, then subscribe to permanent dedicated facilities. However, if you are willing to take a small risk of not getting bandwidth exactly when you need it, B-o-D is a very cost effective alternative to dedicated private lines. Blockage can be experienced since shared pools are not inexhaustible. Check with your suppliers on facility pool sizes, ubiquity, diversity, and blockage rates.

Recovery from Failures
A typical customer-managed private line network has many remote users directed to a single primary enterprise data center. The user via the software’s Graphical User Interface (GUI) manages the permanent dedicated bandwidth between the remote cities and the data centers. Note that the user has a geographically diverse backup data center for use in disaster recovery tests or activations. This could be customer owned or leased. It is more likely to be a subscription to a disaster recovery vendor. In the event of a test or actual site failure, the user can redirect traffic between the remote sites and primary data center to the secondary data center within 3 minutes. Many times these reconfigurations or restorations are within seconds.

In the event of a major geographic incident, the user’s primary data center and the IXC carrier office can be affected. However, the user is able to recover in minutes by individually mapping traffic from each remote city to the secondary location using the GUI of their network management software. Although slightly more expensive than the previous configuration, it demonstrates the robustness and flexibility of these IXC features since none of the remote cities are isolated from the secondary data center.

Frame Relay Service Recovery

There has been limited introduction of frame relay disaster recovery across the IXCs. Those IXCs offering these features are enabling users to complete the reliability picture beyond the inherent frame relay backbone features.

Disaster recovery planning should incorporate scenarios that are created by the customer, IXC carrier sales team, and the IXC technical group with input from a disaster recovery vendor as appropriate. These options cover a number of different recovery situations. Your specific needs and requirements will determine which options are best for your network. This is an effective approach if the recovery scenarios are kept current.

General Features Overview
A key point to remember is that the IXCs do unobtrusive monitoring, surveillance and testing. However, they will not take action until you tell them exactly what you want done. Options are available to protect frame relay networks from access circuit, site, or geographic failures. There are three types of frame relay recovery options:

1. Access protection features allow Permanent Virtual Circuits (PVCs) connected to a failed access circuit to be moved to an alternate access circuit within minutes of a customer declared failure.

2. Backup Permanent Virtual Circuits (BPVCs) allow traffic to be routed from a primary to a secondary location.

3. Growable Permanent Virtual Circuits (GPVCs) not only redirect traffic to a secondary location but increase the bandwidth requirements to meet your application requirements as well. Use of hybrid arrangements of private line and frame relay features allows users to reroute frame relay traffic to other locations, such as a hotsite.

The restoration objective for each of these features is less than or equal to 3 minutes per PVC moved. Total time frames to execute transfers from primary to secondary data sites will depend on the number of PVCs in the disaster recovery scenarios. You may activate multiple disaster recovery scenarios simultaneously to reduce the total timeframe. Re-routing of the PVCs is achieved through use of a specialized disaster recovery software tool in the IXC's frame relay Network Operating Center (NOC). Furthermore, using a specially designed tool, the IXC's network capacity engineers manage and monitor network capacity for customer configurations so that network capacity is always available.

Additionally, you should prioritize which PVCs are to be moved first by pre-defining subsets of PVCs to be transferred to the secondary port. By doing this you can minimize the transfer time of the high priority PVCs that need to be operational more quickly. Therefore, it is recommended that your entire plan be divided into subsets of 20 PVCs or less.

Three Recovery Options
Access Protection Features
Access protection features are best suited for users with mission critical applications/networks that need to recover quickly from failed access channels and/or associated customer premise equipment (CSU/DSU/Router) failures. Specifically, the access protection features are best targeted at customer locations that terminate enough PVCs to justify load balancing the data traffic over two or more access channels and IXC frame relay service ports. Please note that activation of the access protection features could require remote location router table revisions to reflect redirection of PVCs from the primary access port to the secondary access port.

Backup Permanent Virtual Circuits
BPVCs are ideal for users with their own primary and secondary data centers or for users who have arrangements with a disaster recovery vendor to provide backup data center facilities in the event of a site disaster at the customer's primary data center.

BPVCs are targeted at customer networks with key locations such as large primary data centers or locations with mission critical applications. Users can activate or deactivate BPVC scenarios with a phone call to the IXC's frame relay NOC. You must initiate this activity and call to have the PVCs returned to their normal configuration condition.

Please note that only one primary path is active! BPVCs use the same DLCI/IP address at the secondary or backup location as the primary PVCs. This is to enable the redirection of many remote users from the primary data center route to the secondary data center route in a virtually transparent mode. Hence there is no need to revise routers at the customer’s remote locations.

Growable Permanent Virtual Circuits
GPVCs are ideal for users who have multiple data processing sites, or who want to load balance, or who want the comfort of having a hot standby. Under normal (non-failure) conditions, GPVCs are configured with a low Committed Information Rate (CIR). When a failure occurs, the CIR of these secondary PVCs increases to support the increased bandwidth needs to the secondary site.

Once GPVCs are implemented, a customer can call the IXC NOC to activate and deactivate the GPVC scenarios. During activation of GPVCs, the primary PVCs are deactivated and the CIRs of the GPVCs are increased to the CIR of its corresponding primary PVC.

Note that two paths are initially active, hence DLCI/IP addresses of the primary and secondary ports are indeed unique. GPVCs are not as popular as BPVCs since remote router table revisions could be required to reflect redirection of PVCs from the primary to the secondary locations.

Which Technology or Approach is Better for Me?

There are no hard and fast rules. Your decision will vary based on the criteria and assumptions you use. This author suggests the following items should be included in your comparison and decision criteria:

Ease of set up
Set up time
Recovery interval required
Flexibility of use
Number of end points and mileage
Cost

The reason for designing a reliable and secure network is to insure your business is available to your internal and external customers when and where it is needed. The cost of reliability versus the cost of downtime must be weighed against one another. Do an in-depth analysis and create a comparison scorecard to help you decide how much and which features you should pursue. There are a number of options available for ‘hands-on’ and ‘hands-off’ users.

Whichever recovery technology or approach you select, test aggressively and often!!


About the Author
Randy Fisher is Product Manager of AT&T Bandwidth Management (ABM) and AT&T Frame Relay Disaster Recovery Option (DRO) Services. For more information on business recovery for data network management, contact Randy at (908) 234-4655.