Reply to a posting from 10/20/2003:

Hello,

Recently, "David Sworder" <dsworder@noSpam.com> wrote in message 
news:33ejb.42394$th6.30161@twister.socal.rr.com...

>  Thanks for the great information, Doug. Based on your comments, I likely
> won't move forward with this Verizon offering. Perhaps Ricochet would 
be a
> more appropriate wireless Internet service? Do you know if Ricochet offers
> static IPs?

About 9 months ago, a friend of mine in Dallas got interested in Ricochet 
as AT&T (Wireless?) got rid of their test fixed wireless 
telephone/internet service and SWBell had no DSL lines in the area (and 
he either couldn't get or didn't want to pay for cable modem service).

He asked me what I knew about Ricochet and specifically if they offered 
public/static IPs, or if not satic, then were they at least 
addressable/public IPs...

We first called their support number, got no answer, then called to 
Colorado, and they directed us to some number in/near San Jose, and after 
finally getting someone there who admitted that they were planning (then) 
on offering service in Dallas (which they don't seem to be now but did 
mention on their site last winter), they said the only way we could get 
an answer was by filling out on online question form on their web site. 

So I just gave up with them...we still have no idea if they offer static 
or dynamic IPs, nor for that matter if they are even addressable.

>     The fact that the Verizon would offer a service where connectivity
> suddenly drops due to inactivity seems bizarre to me. Does it fix the
> problem if you have a constant 'ping' running in a background DOS box  
(i.e.
> "ping -t verizon.com") to fool the system into thinking that there is
> activity? You shouldn't have to do that, but I'm wondering if it fixes the
> problem.

No and yes, sort of. :)

If we have a telnet (port 23) or Secure Telnet (port 22) session going, 
and there is no data passed *in that session*  for 5 1/4 minutes or so, 
then the connection hangs. Having a ping -t running or even other Telnet 
or SSH sessions with lots of data running at all times makes no 
difference -- the ones with no data passing will hang after 5.25 minutes, 
while the ones with data passing will not.

A running Ping for some reason DOES seem to make a difference in a 
somewhat specific situation:

I noticed that if a SSH session was open, with data coming TO the latop 
but not going out, and if the laptop was mobile (driving, on a train, 
etc., ie, somewhat fast), the connection would "freeze" up UNTIL in a 
separate session you typed something back to the host or started a "ping 
-t" process or in some way you sent out some data (just in general, like 
via HTTP or whatever; not necessarily via the "frozen" session), after 
which the "frozen" connection would unfreeze and whatever data would 
start pouring back in.

Basically, I was using it to review logs, and after about 10 minutes, I 
noticed the logs would freeze. At first I thought it was just a 
connection problem, but (this was a year ago before the idle problem 
started) if I had a second session open which was not passing any data 
and I hit a few keys, all of a sudden the data from the first session 
would "unfreeze" and start pouring in. 

Note that this is the OPPOSITE of how it works with CDPD: If you leave a 
session alone, data will eventually pour into it if you are temporarily 
in a poor coverage area. Hitting ANY key while in a "dead spot" of poor 
coverage will generally kill a session, similar to how a SLIP connection 
seems to work over dialup when their is static over the line. With 1XRTT, 
the opposite seems to be true -- dead spots of poor coverage cause the 
connection to freeze and they will generally not unfreeze unless some 
data is passed out FROM the laptop. This dichotomy between CDPD's 
operation and 1XRTT's was confusing at first, and it took me a while to 
figure out how to tweak our laptops to work well with 1XRTT (an 
especially vexing problem was the Win2000/XP "Media Sense" option, which 
should be disabled with CDPD, but needs to be on for 1XRTT),

(I had been used to and still rely upon CDPD, which unfortunately 
BAMS/Verizon is not building out any longer and according to Rich Wilson, 
a Department Supervisor in Wallingford, will be discontinued it two 
years, when, according to him, Verizon "...is getting rid of all analog 
service". I don't believe him -- I can't see how you can have 
intercarrier roaming without the common denominator of analog -- but I 
can envision that BAMS/GTE want to eventually get rid of CDPD, so I need 
to make sure that we have something decent to replace it, and Express 
Network, at this point, just won't cut it. With CDPD, you never have to 
worry about things like idle circuit drops or running a ping in the 
background to ensure that data pours into the laptop while mobile, etc., 
so these were all new experiences for me when I started using Express 
Network and continued to be such a cause of frustration that when CDPD is 
available, I just use that and really have less grief and aggravation as 
compared to the Express Network Service.)

So other than in the limited circumstance of ensuring a "smooth" server 
--> EN --> laptop data stream when there is no data going the other way, 
no, a continuous ping process will not help, unfortunately :(

> What's the deal with the port blocking? If I understand you correctly,
> Verizon is blocking inbound connections to your laptop, not outbound, 
right?

Correct; certain ports TO my laptop, like 5090 (which I use for a 
java/Web-based "thin client" which allows for remote control of my laptop 
and for people to come in and if not control the machine to at least see 
what I am doing which is useful for remote demonstrations, etc.) are just 
blocked, and I have to spend time figuring out what ports to re-map 
things to to make them work. A real pain, especially when there are 
conflicts with other applications...

NetMeeting is also affected, and I can't seem to get PCAnywhere to work, 
although I think I can get around that by remapping the ports it uses. 

My point, though, is that this is all just a realy pain!

If:

1. They can't fix the idle 5.25 minute problem for Telnet, SSH, and FTP 
(after 3 months!), and
2. They block ports TO the laptop from the outside (and thanks for 
suggesting this clarification, David; I see now my previous post was
    more unclear and incoherent than my usual ramblings :) ), and
3. You need to run a ping routine to ensure that data FROM a server will 
be maintained while mobile (ie, the overall connection can freeze, it 
seems, when 
     mobile, if some data isn't sent back every now and then), and
4. They don't offer a static IP at a reasonable price (and even if they 
did, the port blocking more or less would negate its use), and
5. They have drop problems which can freeze up your laptop's IP 
connectivity...

...then what good is the service anymore? People have wisely suggested 
tunneling and VPN solutions, ping routines, and SSH server options to 
send (invisible/non-displayed) traffic back to my laptop, and I am very 
appreciative of everyone's reponses.

My question to Verizon was (and is): "This 'Express Network' service has 
so many failings and qualifications, why do you even market it to mobile 
professionals? Don't you think they may need to FTP files which take 
longer than 5 minutes? Don't you think they may want to open SSH sessions 
to machines where they may not be authorized to play with the SSH daemon 
to have it send out silent packets just to keep VZW's connection alive? 
Don't you think people might want to use NetMeeting and other 
applications which use presently blocked inbound ports? Don't you think 
they may notice that the connection drops all the time and since there is 
no static IPs available (unless you want to pay $500!) (they do say that 
the IPs, once dropped, are not re-used for a few minutes; on occassion 
I've managed to get the same IP but generally it is a different one and 
all connections thus need to be re-established) all 
transfers/sessions/connections needs to be re-initiated? Not too 
professional, and other than speed-wise, a real step back from CDPD..." .

And to their credit, Verizon agreed, and a number of times said that CDPD 
was probably a better bet for us and others who need the degree of 
connectivity that we do. It's also why we can get out of our contract 
whenever we like (ie, once the Sprint/Sierra 550 card arrives and I see 
if I like it; I know Sprint stinks service-wise and coverage-wise, but 
I'll deal with them if needed because I can't keep going on with 
Verizon's unsolved issues, above.).

Also to their credit, they ARE working on some of these issues: Not a 
week goes by when people at Express Network don't call us up and ask to 
test out the service and see if the idle drops are still happening (they 
just called on Friday, 10/18/2003, and said they found some router which 
they think was causing the problem, so I tried it out in the 00119 
market, the 00022 market, the 00404/00486 market, the 00018 market, and 
the 00028 market, and it is still timing out), and although no progress 
is being made, I get the impression thay they *are* trying to resolve 
this and are making an effort (and I'm sure no other customers  have 
called them about the telnet/SSH/FTP problems, so the fact that they want 
to get it resolved just b/c we reported it and/o).

Indeed, we sent them the "1XRTT Dropped Call List" (see 
http://www.wirelessnotes.org/verizon-and-express-network-drops ) and a 
LOT of those drops have since been fixed (many haven't, but I'd say that 
50% of the ones on the list have been fixed so that the session does not 
need to be re-established or the laptop need to be rebooted). I can't 
imagine any other carrier making an effort to remedy a relatively 
extensive list such as that in under 3 months.

> Perhaps that's why your 'active' FTP sessions aren't working since active
> FTP requires an inbound connection back to the original FTP client. What's
> weird though is that it's only affecting files that take more than 5 
minutes
> to xfer. That is truly weird! What happens with 'passive-mode' FTP?

I think another poster hit it on the head -- if I may crudely re-state 
what he said, the FTP "control" session is what times out, so even though 
there is data passing via the semi-randomly assigned FTP transfer ports, 
after 5 1/4 minutes, the control channel times out and closes the FTP 
session (casuing a number of FTP apps to crash, which is another pain!)

Again, I think it is all related to the idle circuit timeouts which was 
pretty much the "last straw" in times of our patience and willingness to 
put up with the deficiencies in Verizon's Express Network Service, which, 
ufortunately, have not been corrected as of yet.

Overall, I really wish that Verizon would spend the $500,000 or so per 
year or whatever it would take to develop a de minimis Data Quality 
Assurance division which would use their data services (CDPD, EN, Q2N, 
EVDO, MIP, and whatever else they have in store) and drive around, 
testing out the services, thus ensuring INTERNALLY that all this stuff 
worked well rather than making customers test rabbits for these services 
and then not being able to remedy SOME of the issues in a reasonable time 
once these issues are discovered. 

Readers will note on the "Wireless Notes Verizon Dropped Call List" that 
cellular, non-Express Network drops are also covered, some of which are 
*glaringly obvious* (like on I-84 between the 00404/00486 
Orange/Poughkeepsie market and the 00022 NY Metro market), which ANY 
drive test of major roads would reveal have STILL not been remedied for 
years, making their "Can you hear me now?" ads at times seem like a 
hollow joke rather than an indication of any degree of determination to 
fix coverage issues and drops. 

If they are unaware of these drops (which from some of our recent calls 
to them -- unrelated to Express Network -- seem to indicate), then HOW ON 
EARTH can we expect them to monitor the much more technical and esoteric 
aspects of data/IP services to ensure that problems like the ones we are 
currently experiencing with EN will, in the future, be detected 
internally by some apparently lacking Quality Assurance division so that 
we, the customers, are not bothered with these issues and end up getting 
our money's worth for these somewhat highly priced services?

So far, despite their responsiveness to us and their resolving of *some* 
of the dropped call issues, I see no evidence that Verizon Wireless can 
detect, internally report, and resolve some very significant and obvious 
issues with their newer data services (let alone some obvious *voice* 
coverage issues), which, as Verizon is IMO the best of the wireless 
carriers by far, doesn't bode well for the near of medium-term future of 
wireless data. 

Regards,

-Doug

+1 (510) 315-2750
http://www.interpage.net
d and the number 11 with no spaces@interpage.net