ShoreTel News January 2009

Reliability You Can Grow On

Phone systems are transforming business processes with features and functionality that go way beyond simply placing a voice call and into the heart of improving business processes. As organizations move toward unified communications, ensuring the platform is reliable is key to effective communications.

To ensure maximum business continuity, brand credibility and optimum customer service, organizations need 99.999 percent availability from their phone systems. The availability of any phone system, whether a TDM PBX or an IP PBX, is rooted in the reliability of its underlying hardware and software architecture. Here are five core questions to ask when considering whether reliability has been designed into your business phone system — or bolted on as an expensive afterthought.

1. Is call control distributed or centralized?

Distributed call control is a big advantage in phone system reliability. With distributed call control, the switches can be smaller, simpler and ultimately more reliable because there are fewer components and fewer moving parts to fail. In a centralized model, call control is managed through a big chassis server.

Distributed call control is advantageous when it comes to weathering WAN outages as well. A phone system with centralized call control is highly dependent on the WAN connection between office locations, and so if the WAN link goes down, workers in the remote offices won’t have dial tone unless a failback system had been set up in advance. ShoreTel’s distributed architecture puts call control software in the local switch to support basic call control features in the event of a WAN outage.

2. How is redundancy achieved?

No phone system should have a single point of failure, but different phone systems have different architectural approaches to redundancy that have a significant impact on cost and operational efficiency. High availability can be achieved with both N+1 or 1+1 redundancy.

An N+1 system continues to be available even after a single module failure. With N+1 redundancy, the load is distributed across N+1 modules, where N modules are needed to carry the load. If one module should fail, its load is automatically redistributed among the remaining modules.

“Vendors that offer 1+1 redundancy require the customer to purchase twice as much hardware,” said Vijay Gupta, senior product manager at ShoreTel. “With a 1+1 redundancy scheme, redundant switches stand by in case a primary switch fails. ShoreTel offers N+1 redundancy, where each switch is a peer, and the system load is shared across all switches. If a switch fails, its load is redistributed across the remaining switches, and all users can continue to use the system.”

3. What’s demonstrated MTBF?

Disk drives, power supplies and fans are the most unreliable components of any system, and minimizing the use of components with moving parts is essential to reliability. The manufacturer calculates the mean time between failure (MTBF) of its gear by keeping track of the number of units shipped over time, determining how many hours the units have been in service, and dividing that total number by the failed units returned for repair during that time.

“ShoreTel switches have an average demonstrated MTBF of more than 514,000 hours,” Gupta said. “ShoreTel achieves a high MTBF with a switch design that uses high reliability components, such as flash memory instead of disk drives.”

4. What’s the recovery time?

A repair that takes one hour means far less downtime than a repair that takes four hours. The more complicated the system, the longer it takes to identify the problem, replace the failed component and restore communications service. Chassis systems can be complicated to troubleshoot and repair.

With ShoreTel’s modular design, repairs are simple, which results in a lower mean time to repair (MTTR). Installing a switch is a simple matter of plugging in the power source and connecting a couple of cables. The switch configures itself automatically.

Also consider the implication of system design on stocking spares. Complex, chassis-based products require a large inventory of spare components, which is expensive and can quickly become impractical.

5. What about server and application reliability?

People need reliable access to auto-attendants, voice mail, presence and other UC applications. Applications, such as automated attendant, voicemail, application call control interface and call detail reporting, should be distributed throughout the phone system. Unfortunately, many phone systems require that each application have its own server, which adds to the cost and complexity — and can impact system reliability.

ShoreTel distributes voicemail and automated attendant to switches across the network. In the event of a WAN outage, the remote voicemail operates without interruption. Each remote site’s voicemail switch operates independently for the users at the remote site.

If a remote voicemail switch fails, calls automatically route to another switch. The voicemail switch can take messages for users at other locations and push the messages back into their mailboxes when the failed switch is replaced.

ShoreTel uses Double-Take Software for data replication and failover of the ShoreTel HQ server.

Learn More

Having a reliable phone system is prerequisite for business today, but in the event of a business interruption, a reliable phone system can literally save lives. Organizations must consider phone system reliability into their business continuity planning to ensure smooth communication if the worst really does happen.

Download “Building Reliable IP Telephony Systems,” an in-depth whitepaper written by Ed Basart, Chief Technology Officer, ShoreTel.

Learn how ShoreTel proved its resilience when Hurricane Katrina struck Balch & Bingham’s Gulfport law offices.

Learn why Palisades Charter High School put a phone in every classroom for heightened security and improved responsiveness.