The Erlang C model is used for this, it has the built in assumption that if agents are busy then the callers queue. This is much like a Mongrel or FastCGI “agent”, with requests coming in from a front end web server.

Using the Erlang C calculator here, I can figure out how many agents I need, given the desired service level.

• Arrivals: 600/minute (10 req/sec)
• Service time: 0.2s
• Service level: 99.9% within 0.3 s

Gives 5 agents or Mongrels. From there you can work backward with existing models to figure out how many computers you need.

With the calculator you can also fix the number of agents and figure out capacity, or blockage, or whatever.

This assumes that incoming requests follow the Poisson distribution, which I have no idea if it is true. Anyone?
Edit: Sorry, I managed to lose half the post and had to redo it.

a

How many Mongrels or FastCGI processes do I need?

How to configure SIP and NAT
Expose VoIP Problems Using Wireshark

a

My Linux Journal articles are online

How to configure SIP and NAT
Expose VoIP problems with Wireshark

Special thanks to Bill Reid, John Lange, and Les Bester of Les.net for letting me bounce ideas off of them, and again to Les for providing me some testing endpoints.

a

I’m in Linux Journal

A local college radio station is being called to task for the following alleged infractions, committed over a year ago.

The examination also revealed shortfalls relating to the following three conditions of licence:

1. the licensee broadcast a level of 4.73% category 3 music instead of the weekly minimum of 5%;
2. the licensee broadcast a weekly level of 0.83% news instead of the weekly minimum of 4%; and
3. the absence of any formal educational programming despite its condition of licence to broadcast a minimum of two hours of such programming.

A hearing was called and people were flown in from all over the country to attend. All because a college radio station didn’t play some news and missed a few minutes of “category 3 music”.

I realize spectrum is a limited commodity, but gee-whiz, isn’t there a more efficient way of handling this?

a

My tax dollars go where?

It’s actually quite easy in theory, since you can Dial() any sort of address, but the trick is to integrate it with the dialplan.

A bit of research found this page which is good, but it assumes that everything gets forwarded to the SIP macro. Using part of his recipe:

exten => _[a-z].,1,Macro(uridial,${EXTEN}@${SIPDOMAIN})
exten => _[A-Z].,1,Macro(uridial,${EXTEN}@${SIPDOMAIN})

means that only those addresses that start with a letter will be dialed as SIP, which rules out numeric addresses (like 613@fwd.pulver.com for echo testing).

I like his idea, though, so I decided that I’d keep it, and prefix any numeric sip addresses with sip: so that the extension above would be caught. Then it was a matter of modifying his macro to look for sip: in front of a URI and strip it. The results are:

; Dials a SIP URI
[macro-uridial]

; First, assume it’s not a sip: type address
exten => s,1,Set(destination=${ARG1})
exten => s,n,NoOp(First 6 chars are ${ARG1:0:6})
; We see them as url encoded (ie : is really %3a)
; check to see if the first *six* chars match it then
exten => s,n,GotoIf($["${ARG1:0:6}" != "sip%3a"]?nowdial|1)
; we fell through because there was a sip: at the beginning, so
; strip the digits and then dial
exten => s,n,Set(destination=${ARG1:6})
exten => s,n,GoTo(nowdial,1)
; At this point it’s a proper uri
exten => nowdial,1,NoOp(Calling remote SIP peer ${destination})
exten => nowdial,n,Dial(SIP/${destination},120,tr)
exten => nowdial,n,Congestion()

In the end it turned out to be an exercise in learning how Asterisk works more than anything else, ie the %3a translation, the implicit breaking up of the address into $EXTEN and $SIPDOMAIN, and also the format of GotoIf.

a

Asterisk and dialing URIs

ENUM is a method for mapping PSTN style phone numbers to a URI, such as SIP or a mailto, through the global DNS. My office phone, 204.975.5987, would then become 7.8.9.5.5.7.9.4.0.2.1.e164.arpa by reversing the digits, adding the top level country code of 1, and looking within the e164.arpa zone. Simplifying greatly, the response might either be nothing, meaning I can’t be reached by IP, or it might be something like sip:seanwalberg@ceridian.ca, meaning that I can be reached via SIP through that address.

One consumer of this information would be telecos. Rather than paying my carrier (MTS/Allstream) for terminating the call, they could use the sip URI to call me over the Internet and pocket the difference. The other would be regular people who are replacing their phones with either computers or smarter IP phones, and want to make the query themselves to avoid paying long distance.

Almost two years ago I wrote Either you’ve got a SIP address, or you’re not worth talking to to point out that the eventual goal will be phasing out of PSTN style numbers in favour of SIP URIs, which are basically email addresses.

Since telephones with their 12 buttons aren’t too good at entering SIP URIs such as “seanwalberg@ceridian.ca”, ENUM is one transition mechanism. But involve telecos and large committees, and you get delay and disagreements. Already there are two types of ENUM being proposed, one for carriers, one for the public. And since DNS is delegated differently than PSTN numbers, how does a regular person like me update my records securely (while making sure that an evil person can’t update my records?)

So this led me to think, “Why do we even need ENUM?”. Why not put a SIP address on your business card along with all the other numbers people put there? (Don’t quote me, but I remember Nortel did this) People who want to use SIP will use SIP. People who can’t won’t.

From the carrier perspective, if they’re so interested in saving money by offloading calls to the Internet, why not let them come up with their own transition mechanism? They’ve done this before for toll free and local number portability.

I suppose what I’m really saying is that the problem isn’t so much technical as it is marketing. How many people have some form of instant messaging client, or even several? What would it take people to move to a SIP based client? If people started using SIP based instant messaging, or even if it were supported under popular IM clients (didn’t MSN support it for a brief period), wouldn’t vendors have incentive to make hard phones capable of dialing SIP? (I pointed out last night that a small keypad was no impediment to the adoption of SMS messaging)

While I applaud the efforts of those involved with ENUM, I wonder if it wouldn’t be easier just to draw a line in the sand and not try and integrate the two networks. Instead, focus on promoting the SIP network, and let the benefits and features draw people in.

a

Do we really need ENUM?

On another voice related note, John Lange has started up the Canadian Association of Voice over IP Providers, a sort of lobby/working group for Canadian ITSPs. The goal is to present an independent perspective on VoIP regulatory issues to the Canadian Radio and Telecommunications Committee (CRTC) and to work in cooperation with Local Exchange Carriers (LECs) to ensure interoperability between VoIP networks and the Public Switched Telephone Network (PSTN).

a

Telephony Terms and Concepts Article

It’s an interesting thought, with many good arguments for and against.

On the surface, it may appear that BS is trying to have their cake and eat it too. People using their network pay for the access already, either by cash or reciprocal traffic arrangements. So, company X may connect to provider Y who in turn connects to BS. Y pays BS for the traffic it uses and X pays Y. BS may connect to another ISP who agrees to a reciprocal arrangement whereby each carrier will carry the other’s traffic free of charge.

In this scenario, the network is paid for, and any argument about “the Internet is getting faster and we can’t keep up” should be able to be taken care of by BS charging more to Y who passes it on to X should X be the person using the traffic.

On the other hand, one could look upon this as a value-add. “Yahoo!, pay me, and I’ll make sure your traffic has priority on my network”. (As an aside, giving traffic priority doesn’t always refer to a strict priority queue where Yahoo! always beats out Google and other traffic. It could be a guaranteed portion of the pipe) From this perspective, it seems reasonable to auction off different priorities within the BS network, assuming there is a market.

Of course, the “value” part of “value-add” is dubious. BS isn’t the whole Internet, and the hop-by-hop nature of the Internet means that your traffic will only experience the increased performance over the BS network. If your cable modem network is pinned, this won’t help you.

Now that I write it out, I’m starting to think the free market approach is better. One of several things will happen:

- People will pay. By this, I mean the ultimate end product is priced such that people buy it, and the content provider maintains a satisfactory profit
- People won’t pay. If I’m only willing to spend $1 on a music download, and I won’t spend $1.05 to have it download 10 seconds faster, then no one will choose the high quality option, and we’re back to the Internet we have today where all traffic is equal
- It won’t work for the reasons cited before. Other carriers could choose to alter their peering to avoid the BS network, or the benefit of the service could be so small that people won’t pay.

Either way, I don’t see a long term problem. The market should take care of it, whether or not it’s good for BS is the question.

All that said, the old adage often accredited to John Gilmore applies: “The Internet treats censorship as damage, and routes around it”. In this case, if BS tries to push it too far, or misuses the idea, the Internet will take care of itself.

(Also have a look over at the VoIP blog for another take)

a

Bell South, and charging for premium service

I must say it was easy to get signed up. I filled out a web page, and a little while later, I was sent a package containing my VoIP box (ATA) and instructions. I soon got an email with the day my number would be ported, and it was transferred when they said it would be.

However, quality problems made the service unusable. I played with QoS, called Tech support, tweaked this, tweaked that, and eventually gave up. This, I don’t fault them for. I have a few friends who are happy with the service, it just didn’t work for me.

So, I called up Shaw for their Digital Voice package. They gave me a temporary number to use, I set up my TalkBroadband line to forward all calls to that number, and disconnected the ATA.

I then called Primus customer service to switch to the lowest plan possible, $15/month (if I’m not using the service, why pay $30?). Surprisingly, I wasn’t asked why I was leaving. Even after working in “I’m just not happy with the service”, I didn’t get “sorry to hear that”.

An unfortunate mixup with Shaw caused the number porting to take much longer than it should have. (Other than that, I’m very impressed with Shaw’s service. The guys that came to wire my house did an outstanding job) Any calls to my Primus number which were call forwarded to Shaw (never touching VoIP) had awful quality, such that we usually ended up calling people back whenever they called us.

So, I called Primus support. The first guy I talked to seemed rather angry at me and very argumentative. I called back through customer support to complain, who had no record of the previous call. So, I got someone else in tech support, who blamed the problem on Shaw.

Around the same time, when people called us in the evenings, they’d sometimes get an “out of trunks” message. I called Primus about that, too, and the support guy told me that he could only help me if I was getting the message now, and to call back when I was having the problem.

Finally, on Dec 21, my number ported over to Shaw. I was told my account would be cancelled. Late last week, a friend on TalkBroadband tried to call me and got a “not available message”.

So, I called Primus this morning. The lady told me my account hadn’t been cancelled, and since I have to give 30 days notice, they were billing me until Feb 9. Fine. I then asked her what to do about the problem where Primus customers couldn’t call me and she told me that I couldn’t possibly be having this problem, and that the person calling me should call in. After trying to explain my situation to her, I asked to be transferred to tech support.

The guy on the tech side was nice enough, and seemed to have an idea of what might be wrong. Since my account wasn’t cancelled, Primus customers were still going the old way. He called some people internally and told me I should wait a few hours for the account to be cancelled inside their system first.

All told, I’ve spent around $50 on a service I didn’t use (that’s not including the $30 on the month of service that I did use), and have had many problems dealing with Primus’ so called “customer service”. I have never dealt with so many rude people as a customer of a company or felt less valued as a customer.

I have friends who are happy on Primus, though they have never had to call for help. (That said, one of them did have a billing problem, and it took a very long time to resolve because of the pathetic customer service).

On the flip side, Shaw has been nothing but helpful. I pay more for the Shaw service, but it is excellent, and I feel taken care of.

a

Primus TalkBroadband sucks

Googling around, I couldn’t find a good example of anyone who has done it, so I was on my own.

My goals here were simple:

1. People dialing 555-1111 should be picked up by the AS5350’s internal modems
2. People dialing 555-222X should be forwarded to my Asterisk box for handling
3. The Asterisk box should be able to dial out using the PRI

Already the DIDs were assigned to this PRI, so that wasn’t a problem.

The configuration on the AS5350 is fairly straightforward once you get it working:

voice service voip
 sip
  ! Bind the SIP and RTP sessions to a known interface
  bind control source-interface FastEthernet0/0
  bind media source-interface FastEthernet0/0
! Normal T1-PRI configuration using all channels
controller T1 3/0
 framing esf
 linecode b8zs
 cablelength short 133
 pri-group timeslots 1-24
! Anything not picked up by a dial peer defaults to a modem call
interface Serial3/0:23
 isdn incoming-voice modem
! Use a data dial-peer to answer anything to 555-1111 with a modem
dial-peer data 10 pots
 incoming called-number 5551111
! Match inbound calls to the voice DIDs
dial-peer voice 100 pots
 incoming called-number 555222.
 ! All digits are collected and sent along
 direct-inward-dial
 forward-digits extra
! The IP leg of the call... send to the sip server defined below
dial-peer voice 200 voip
 destination-pattern 555222.
 session protocol sipv2
 session target sip-server
 ! Any digits collected after the call has been answers get forwarded
 dtmf-relay sip-notify rtp-nte
 ! I want g711 ulaw for a codec
 codec g711ulaw
! Anything else gets sent out the PSTN
! Note you need the data dial-peer above to accept modem calls, otherwise
! this dial-peer will give dialtone to anyone who calls!
dial-peer voice 1000 pots
 application session
 destination-pattern .T
 port 3/0:D
 forward-digits all
! define your SIP server
sip-ua
 authentication username USERNAME password PASSWORD
 ! xxx.xxx.xxx.xxx represents the Asterisk box
 registrar ipv4:xxx.xxx.xxx.xxx expires 3600
 sip-server ipv4:xxx.xxx.xxx.xxx

After that, it’s just a matter of configuring a sip peer in /etc/asterisk/sip.conf

a

Asterisk and an AS5350 SIP peer