Phone system for multi-office company

We are currently looking into replacing all the phone systems at our health centers.  Over the past few weeks i have been in talks and have seen demonstrations by the different vendors, but now i would like some input from the community here.

Suppliers in the running: Avaya, Cisco, Shoretel, NEC, Toshiba

Situation: Health Centers that provide out-patient care.  Has fiber WAN with 10Mbs connections to each site(180Mbs WAN metro-e circuit), 15 locations, 400 handsets.  

Wants: Move to one phone system for all the centers, use the fiber SIP trunks offered by the telco at main location and then pipe all calls through main location.  Have 911 services local to each location via analog lines( up to 4 lines ).  Have failover via the WAN, if the local phone system dies/blows up/whatever, the IP phones at the site connect to another box and continue to function.  If the WAN connection dies, we need the local system to use the analog lines for outbound temporarily.  If the Main site SIP lines go down, the ability to use another SIP or PRI at another location for the entire system would be great.  We are looking for easy provisioning, good reporting and the ability of staff to make changes to their extension options as well as supervisors to their site extensions.

All the suppliers above want to provide us with a solution and from what i can tell, Cisco, NEC and Shoretel can provide it.  Avaya seems slightly behind in features/fail-over and i havent seen the toshiba system yet, but i can say we have 2 toshiba phone systems that are 20+ years old that are still running.  Cisco looks very expensive but seems very serious about delivering.  NEC seems to have everything we need and more after i install 5+ servers.  Shoretel seems to be the easiest to setup and has the reliability out of the box we want at a much lower cost than cisco.

Any thoughts or experiences with each one ( use, call quality, setup, maintenance, costs, etc)?  If you are going to bash some.. please provide real world examples of failures and whether or not you think it was the installer and/or the manufacturer at fault.  I am looking for as much detail as possible and if you want to vent.. thats fine... examples make it more legitimate for me though.

Thanks Ahead!!!!
Who is Participating?
Joel_SiskoConnect With a Mentor Commented:
The issue at hand is not the individual phone system themselves. They all provide dial tone and are reputable companies. The issue is defining what is most important to you based on the features sets that you are looking to implement.

Also have you looked at open source solutions such as Asterisk? They may present more flexibility and at lower price points in most cases while still using name brand phones that the other systems support as well. I preference my last comment on the fact I design and install open source solutions, prior I was installing Avaya, Nortel, Cisco and Toshiba phone systems for the past 16+ years.

Now you mention the vertical, Health Care, which in my experience has  meant a whole bunch of analog handsets as extensions, this poses a problem with many of the these systems since they have limitations on the number of types of sets that can be supported from a single system. Therefore your design may require more that one system within the network.
Also as a personal preference and based on experience, PRI circuits are the best way to go, the most reliable, do not care who the carrier is, I have yet, to find a carrier that can deliver reliable based VoIP circuits to the premise (I watched Verizon take out all of its VoIP clients last year for a period of a few hours in the Southern California market, with good old reliable Cisco gear at the helm).

If I understand these are your requirements,I  listed systems in order as being best to worst, left to right:

Redundancy at the PBX level- Cisco, Shoretel, NEC, Toshiba, Avaya
Redundancy at the Carrier level- Shoretel, Cisco, Avaya, NEC, Toshiba
Ease of Administration- Toshiba, Avaya, Shoretel, NEC, Cisco
Grand unification of dial plans- Shoretel, Avaya, Cisco, Toshiba, NEC
User Portal- Shoretel, Avaya, Toshiba, NEC, Cisco
Cost of ownership- Avaya, Toshiba, NEC, Shoretel, Cisco

Kindest regards,

Joel Sisko
I really enjoy working with the Cisco solution. I've never used any other VoIP system besides theres. It can, and does everything you are asking for (and more) very well.

There is a large learning curve but think of it as a chance for you to go to training :)

Your scenario is what we have at my work place. We have 5 PRI's at main location with 4 POTS lines at each remote location (40 of them!).

The thing i would be most worried about with the Cisco deployment (as of most deployments) is whether or not you have the infrastructure to support a VoIP deployment (Namely QoS, PoE capable switches)
Hi there

Cisco solution i will recomend , lots of support and easy to configure.
Improve Your Query Performance Tuning

In this FREE six-day email course, you'll learn from Janis Griffin, Database Performance Evangelist. She'll teach 12 steps that you can use to optimize your queries as much as possible and see measurable results in your work. Get started today!

BrentDevOpsAuthor Commented:
Thank you jjmartineziii and Joel_Sisko.  Greekstones.. could you elaborate more on your experiences with cisco?
Joel_Sisko: We are planning on deploying all VOIP handsets, no analog or digital handsets.
 jjmartineziii: QoS is a big worry for me.  We are currently replacing our dell 10/100 dummy switches with HP 2510 managed units that have very basic QoS controls.  I am actually going to halt deployment of those because i am assuming we will need to buy PoE switches or injectors and i want to have the ability to move them around as needed without having a large supply of 2510's lying around because replaced them with PoE.  It may be necessary for us to replace them with more robust QoS switches or at the least have one powerful QoS switch at each site that maintains the flow accross the WAN.
JoeL: I like your layout there.  I have been leaning towards shoretel mainly because of their reliablity and recovery options.   As far as open source goes, i have played with asterisk and trixbox and free switch.  They all seem to have the features i need in general.. i just dont see much in the way of universal fail-over and almost nothing for call-center setup.  In the call center, a manger is going to make sure everything is on track, they need an easy to use interface and i cannot seem to locate anything useful in the open-source world that works with trixbox or free-switch.  Thoughts?
jjmartineziiiConnect With a Mentor Commented:
Yes, also, your routers at the remote sites need to be capable of handling SRST in order to have failover! Another consideration if you plan to use Cisco.
greekstonesConnect With a Mentor Commented:
Hi Icetoad

I work for a Cisco Parner and Avaya Partner ,  My experince in voip is around 10 years and i find Cisco more stable than other voice products out there. Also if you go for Cisco Telephone Maintance is much cheaper than any other products.

FreeSwitch is not the right solution at was designed and started to be a soft switch replacement and application platform for developers. You need a PBX, Asterisk was built to be a PBX. There is a huge difference between the two.

Asterisk can be used in a very reliable fail-over solution. We provide an Asterisk solution that uses two servers in a cluster and single PRI circuit which automatically fails from the Primary to Secondary server in 10 seconds. This includes switching of a single PRI between the two systems automatically. Then we provide continuity with our Carrier partner and route all calls to a separate site if the primary site should physical go down automatically.

We use Vicidial for all our call center applications most recently with a client in Canada that handles 15,000 calls per day across multi-sites with agents using a variety of custom applications which we integrate the phone system with. Do you have any custom applications that you need to work with in regards to your call center? Any IVR/Voice Recognition needs?

Do not want to get too much into a marketing pitch (but always glad to do so). If I had to choose between the solutions thus far you have looked at: Shoretel would be my first with Toshiba in second.

Joel Sisko
BrentDevOpsAuthor Commented:
The main concern i have with asterisk is the ease of use.  If i were to leave,hit by a bus, etc..  just after installing it; would someone be able to step in and continue servicing it?
As far as redundancy is concerned... it runs on a moving hard drive.. which to me.. has more potential to fail and takes up more space than an a hardware appliance.  Obviously there are some options that still need a server no matter who you have.
My other concern is phone provisioning... paging, intercom, buttons, free/busy... which handsets work best?  I have a 7960, an astra and a 2020 grandstream that i tried with a trixbox config.  I was impressed by the grandstream and the cisco.  I hear polycom is also a great choice.  How would i mesh different handsets and still maintain a good feature set?
I do like many of the customizable things in asterisk.  I was playing with cepstral text to voice.  Streaming online radio stations, conference rooms, calling queues.  
The other issue is finding a solid firm in my area that can handle this.
Where are you located?  What system(s) do you have in place today?

LIE theory I use all the time, if you were hit by a bus on the Long Island Expressway (LIE) would the business continue to run. The answer is it is the same answer when you use a Cisco, Avaya, NEC, Shoretel or Asterisk. If you were the only person trained on the system, the knowledge goes with you. Going to be a bit facetious here to make a point about a few items in regards to GUIs and ease of administration.

The question is can I get someone external to the company that can handle this for me. The question is yes to all the systems available. Mind you, Asterisk based systems are now 18% of all systems going out the door in new IP PBX sales. Remember these are networked systems, both LAN and WAN, therefore, as in our case we service all of North America and parts of EU with only four offices.

Also you are trying to put in a single system that is going to create a complex voice network correct? If so, how many people do you have trained for the complex data network today? Do they have a simple to use interface to administer the entire data system? Do you have an interface to manage your cell phone? How many times do you change the network configuration of your cell phone with your provider? You do not, yet you use it everyday. So why must you have an interface for an extremely complex phone system that enables someone that has zero knowledge about phones and communications be allowed to manage it? Do you do the same for the data network?

What you do need, IMHO, is an interface that allows users to change things on a whim, i.e. what number to forward my office phone this afternoon. An interface that allows managers to add new users and even more important lock out fired employees without having to spend 10 years in the field installing systems or go through a CCIE course from Cisco.

Now regards to phones, hands down Cisco makes the best, but they also cost the most. Grandstream and Snom we stay away from because the automated provisioning of these phones is just not easy. We use Aastra and Polycom as preferred phones, followed by Cisco. From a remote service perspective Aastra wins hands down over Polycom. Plus our ability to integrate custom applications with Aastras is much easier than that of Ciscos.  

Regarding hard drives, all depends on how the solution is designed. As in our case we use Solid State drives (140 year MTBF) in our Asterisk platform. All systems though, use the same CPUs and other components. Also note if the hardware is designed specific, so is the software, therefore flexibility in adding new features is very limited (i.e. Shoretel and Cisco).

Joel Sisko
BrentDevOpsAuthor Commented:
We are located in orlando FL.  My worrying with asterisk has to do more with the customizations that a possible.  If i tap in and enable some custom scripts to perform text to voice or stream radio stations or rotate mp3 files, provide weather reports or stock tickers.. how is that knowledge going to be transferred to someone trying to fix it especially if no one else in house has training?  I would want a local VAR that could handle it when i cant or provide those types of customizations.
We woud like a single unified system, but it must be reliable, have good quality and be able to recover when a gateway or connection fails.
You're dead on with the management.. we want web or thin type management for admins here and then web based local extension management for site managers.  They also want the ability for users to do the simple stuff with their own extensions.  They really like the HUD's that each company seems to offer and the presence features just about sold them on the unification of each system.
Currently we have 10 locations and a different PBX at each.  We have NEC, Voldalvi and Toshiba systems out there.

Are all 10 locations in FL? Are they each serviced from a carrier perspective using analog or T1/PRI circuits?

Whether you do a custom stock ticker script on a Cisco system, Shortel or Asterisk you run into the same issue-- it is custom to that system/manufacturer platform. You can not take a script from Cisco system and place it on a Shoretel system. Also do not assume just because it is Cisco or Avaya another Cisco or Avaya VAR can work on it. In some cases Avaya partners can not work on a different series of equipment without risk of loosing their VAR dealerships. The point I am driving home about Asterisk is that you are correct you need a competent VAR. You have good and bad VARs regardless of the product line.

Asterisk is worth a look at not just from Trixbox perspective, but from companies that are using Asterisk and other open source products to deliver reliable solutions. We have several clients that started down the Cisco path, but are now using Asterisk because of its ease of use and pricing. BTW, Cisco, Shoretel and Avaya use open source in their current IP offerings.

Joel Sisko
BrentDevOpsAuthor Commented:
Yes, 8 locations in one county and 2 location in the next county over.  The fiber circuits ( Metro-E ) for data are all serviced by the same utility company.  The current phone circuits are all over the place.. some are PRI, some are analog lines.. providers vary.. paetec, embarq, att, sprint.  
We want to use our existing data provider who can provide SIP trunks directly to the new phone system.  then maybe a backup PRI at a different location to ensure the sites dont completely rely on the main location.. just incase the lines get cut.
I agree on the VAR issue.  It seems many of the complaints i see on the web seem to be related to VAR implementations, not hardware.
I was originally looking at asterisk on a per site basis, then they decided to go all out with a phone system.  I just dont have enough experience to peice together a solution to our new needs in a reasonable amount of time.  I actually seutp a trixbox system and a admins paradise system and attached a cisco 7960, astra 9110, grandstream 2020 phone to them to see which handsets would provide what we needed.  The phone system needed a little chat help to get an outside trunk setup on vonage for testing.. but after that, worked perfectly.  Another concern i had was how much call volume it could handle.. being that we want to setup a call center.  Of course to me, i see the system as unified, but most VAR's seem to place call centers in their own little sub group.
When do you need the system in place? Having a PRI as the backup would be good and should be a requirement in that size network. What do you need in regards to your call center needs?

Joel Sisko
Just to address the capacity issue with Asterisk. When using PRI with a Quad Core CPU 2.66, you can get 4 PRI at full capacity and around 1000 SIP endpoints. When using SIP trunking, then you can have more carrier connections than the PRI's. Hope all this helps.

Joel Sisko
Ron MalmsteadConnect With a Mentor Information Services ManagerCommented:
Having used both Cisco and Asterisk, I prefer Asterisk for two main reasons....
1) cost is a fraction of what Cisco costs to build and maintain.  Requires less hardware which is the main factor in reducing cost.  No licensing fees, lease contract agreements, etc.

2) customization and functionality...   simply put...there is nothing you can do with Cisco that you can't do with Asterisk as long as you have the creative ideas and technical resources.

If you prefer "dummed down" versions of Asterisk like AsteriskNOW, or business class supported versions....then the adminstration learning curve can be offset.

The last place I worked for spend nearly a million dollars for an enterprise Cisco installlation that took months to finish... there was a huge administration learning curve because of AD, exchange mailbox, and RightFax integration,.....but at my new job,... I was able to put more functionality into a single Asterisk box to support the same amount of users, with a shoe string budget, in less time.....and we don't have to make payments to anyone...we own it.
Ron MalmsteadInformation Services ManagerCommented:
btw- If someone is "hit by a bus"..... you'll want to have documentation on the setup and configuration, regardless of whether you use Cisco or Asterisk.

Cisco voip installations can be very complicated depending on the needs of the business....just as much if not moresoe than any Asterisk implementation.
BrentDevOpsAuthor Commented:
Joel: We would like to complete the project by the end of the year.  Start time can be anytime in the next two months.  The call center is about 20 stations.  Currently we have an issue with permits for a permanent connection to that building.. but hopefully i can work that out before we start or during the setup.  We are looking at SIP trunking directly from the carrier.. total of 96 channels.
XUserx2000:  I realize the "hit by a bus" scenario is generally moot with the right documentation, however, i am the only one at my office who knows *nix.  Its not something you can train most people overnight in.  As far as the customization part, i have seen that as i have done some myself on a test box.  The issue again comes back to training another person on how to do so.  If they dont understand linux.. its harder to teach them how to modify scripts.
I believe you on the cost issues.  I also like the ability to have different handset from different manufacturers.  I had a test setup with a cisco, astra and grandstream handsets, it worked with all three.  The only issue i had was provisions each phone properly and setting up certain settings so they acted the same with each handset.  I am sure that when we deploy, we will stick with the same brand to save money and ensure standardization.  My friend handed me over to a VAR out of NY who has a VAR in FL, so hopefully they can give me some nice $$ that my boss will like.  Though with some recent funding changes.. we may be able to stick with a full services company.  
I guess asterisk makes me nervous.....  If we invest all this money and it doesnt work or doesnt work the way they want... things could get ugly.  Though on the bright side.. it would probably still be better than what we have.  With that said, has anyone made a heavy duty appliance(no moving parts or maybe one fan) that runs asterisk that can handle our load?
Ron MalmsteadInformation Services ManagerCommented:
I'm located in SWFL, and I work for a telmarketing company that uses Asterisk for our 150 seat call center.

I have a Dual core Dell tower running Centos4 and Asterisk installed/configured..... it has one FXS and one FXO analog port, one T1 PRI port, and can accomodate sip trunking for 20+ users very easily...  I never got the chance to "max it out" however,...

This box served as a 2 week, temporary solution to building a bigger server that we now use here.  If you would like to demo it, we could probably work something out.

I would be more nervous about investing in Cisco, since it costs so much money to implement and maintain.  If Asterisk turned out to be something you dont' like.... you really only invested money in the hardware and time,...and the hardware can be recycled for other needs or uses at that point.  You also now have the option to purchase a support agreement for all versions of Asterisk through digium, which was previously a "one up" for Cisco who always had a large multi-national support group.

If Cisco turned out to be something you didn't want....then chances are you are stuck with it, because you have probably already saddled your business with a hefty proprietary hardware investment, and long term licensing and leasing agreements.

I think the reason that Cisco is more mainstream for voip, thus giving people the "feel good" factor... is this....  there is more money to be saved by using Asterisk than there is to be made selling it as a solution.  As a voip systems installer, I could quote you a $500,000 implementation for Cisco, or I could quote you a $5,000 implementation for Asterisk...even though both of them do the same thing...which am I more likely to try to sell you ?  Cisco duh !.....and i'm going to try to convince you that there is no alternative that can meet your needs even though there are many.  Of course after talking up how easy it is to use, and how reliable it is..., i'll still try to convince you to pre pay for an onsite maintenance agreement.

If you are curios in seeing a quote for an Asterisk solution I can be contacted via my email address in my profile, but want to make sure I answer your original question. With the systems that you have mentioned one thing to keep in mind is the licensing, Cisco can become a nightmare, followed by Shoretel, with Toshiba being the easiest to deal with out of the bunch. Toshiba offers the most features and easiest to manage of the three, but will lack the ability network easily across the setup you have. If you consider separate small system for each site an then a unified dial plan, Toshiba can be a great solution.

In regards to moving parts to non-moving parts, is well&marketing in most cases by Shoretel and the like. There is validity in you get what you pay for with electronics, so as in our systems we sell a Intel chip in a platform we engineered for about twice the price that you can get from Dell. The difference is not in the Intel chip, its the components around it. We engineer our systems as telecom systems, not web servers and have revision control minimum five years with up to eight years. Therefore I can support your platform five years from now. Ask Dell for the same server you bought five years ago, let me know the response. When dealing with a single sight solution it is not an issue, but when you have hundreds of clients or clients with hundreds of sites, the ability to provide the platform they purchased five years ago is crucial for a multitude of reasons, mostly support and engineering.

Ultimately your fear is making the right decision because that is your job. The issue at hand is not the technology, the decision is in choosing the right folks that can evaluate your current setup, evaluate your business needs; and then find the right system to deploy based on those needs. Implement that system correctly the first time, train the users and support those users after the post cut. Then train your staff to be self sufficient on the system without having to be Linux gurus.

Joel  Sisko
BrentDevOpsAuthor Commented:
Thanks for the comments guys.  I have enough resellers to talk to right now and i am speaking with a SipX vendor.  I have played with asterisk but i fear i will not have the time to build and test a system that does exactly what we need and out of the box.
As this point i really need to concentrate on the call routing for each location, the call center, hunt groups, number porting.  I was approaching this from a leisurely perspective and my boss just lit a fire under my ass to get it done yesterday.  I really hate rushing stuff like this and this was not in the cards originally.
I am going to go ahead and award points for your efforts.  Thanks for the input!
BrentDevOpsAuthor Commented:
Joel and Xuser, i really appreciated your input, points are split accordingly.
BrentDevOpsAuthor Commented:
Just incase anyone wanted to know, we ended up purchasing a shoretel system with the call center.  Easy setup, the call center was the hardest part.  The actual site phone setups and change out were just so simple.
We have had some RMA phones, about 10 so far out of the 350 we bought, so not too bad.  The phones themselves can take a beating though.
The users love the phones and the PCM client.  I have never in the 9 years working here had anyone call me excited about some software we installed... the PCM client gets me emails and messages with people just extactic with regards to all the features.  They also love the login/logout feature, but i think thats pretty standard for VoIP systems.
We ended up with three phsyical servers: Conferencing Bridge, ECC contact call center, and Director.  all of which are single points of failure.  Luckily they are working on the virtualization part, so when 10.1 comes out, we will probably upgrade and deploy that for redundancy.
The hardest part was the call center.  It is easy to use but there are so many settings, its quite difficult to track down exactly what you need to set from scratch.  The reporting part is hard for end users to understand, so make sure they get educated.
The director reports are lacking.  I wrote my own mysql queries to pull data as it was requested by upper management.  I am not happy with that, but there are add-ons you can buy if you dont want to go my route.  It just was not in our budget.
Cisco was not much more but i am glad i didnt go that route.  I did some side work for another company who had used them and they said they were a terrible VAR ( not cisco itself though ).   They are still working with cisco directly to get their system fully functional.
The only really negative thing against shoretel is that they require you to talk the VAR only.  They will not talk to you directly unless you buy an enterprise support contract( i might do this, though i dont think we are big enough ).  I have to rely on our VAR to get answers.  Sometimes they are on the ball or know the answer and sometimes its weeks before i get a final answer.  That may or may not be a product of my questions and weird stuff we do.
I hope this was a fair assessment.  Overall our company loves the product and not once has anyone mentioned trying to rip it out!  ( which was my one of my goals ).
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.