6 replies to this topic

[cybermaus]

Hi

I got the new explanation for the split of SPLASH into FREE and FREE premium. That is a bit worrysome.
I can understand the drive behind it, and in fact, while I like some of the now 'premium' (former free) items, (especially the router down notification), if I must I can do without.

But the footnote at the bottom is the one that maybe you want to reconsider: The fair use policy.
Only 500 sessions per month? That is only 16 per day?

You understand, we try to run a free service, because we have such a low backbone we could not possibly charge for it. So it is just a courtesy to the visitors to check their email and browse (non-media) web sites. Its also why (hithirto) we plan to log them off after one hour, so one hour sessions only. This so phones and laptops left running are not constantly polling.

16 per day is impossibly low though. It will be used up on kids alone, with their phones, who were told they cannot use youtube, but will try anyway.

Can I suggest you reconsider the sessions per months? Even a tenfold would still be a low number.

Many thanks

[Roland]

Hello,

i can tell you from our statistics that 500 sessions is enough for about 80% of basic locations and that's still a huge number.

Are you really telling that you cannot afford $4.90 per month for the service? That's about a price of a coffee. We don't try to make huge money from this. The money we can get from this would hardly be enough to cover our costs (support, constant improvements, hosting and operating servers, etc). If you like the service we provide why is it so hard to support us with such a small monthly fee?

I would like to hear your opinion about this, maybe we are missing something here...

[cybermaus]

OK, I have upgraded and paid 6 months' fee.

I still have one worry though. If I look at the location summery page, it is listing:  

Today : 0 Transactions; 34 sessions
This month: 3 transactions; 818 sessions

Transaction wise, this is correct. The campground is not yet open, and only a few (could be 3) test sessions were made.
But why is it listing 818 sessions, and 34 sessions for today, with no-one on-site, and no transactions.

In the past, I did not care, but right now, I am worried. Even as a plus user, I only have 5000 sessions, so what exactly is a session, how do they relate to transactions, and why do I already have 818 with no-one on site?

Many thanks for your help?


EDIT:

Digging through the session report, these are all almost all my own network devices. The devices are pre-authorized in Coova (HS_MACALLOW) and on a non hotspot IP range. So even though these devices are not going through your servers for logon or radius. but they are still counted whenever they do a NTP time sync, or when they report they are still alive.

There is also a pre-authorised Samsung Android. It also does not go via your server, but it seems that whenever she walked in and out of range, it is counted as another session.


So, I can understand your argument about low-fee. Which is why I paid up.
But you also asked me if you missed something, and I think this may be it. Possibly you should either filter out these non-sessions, or set your fair-use based on transactions rather than sessions?

Many thanks

Edited by cybermaus, 15 March 2012 - 02:24 PM.

[Roland]

Ok we will check this next week, but it seems that those sessions are still registered in our servers as coova is sending the accounting information to our servers, even though they are local users (i guess you are using local macauth, right?)

From the system's view, these sessions give our servers the same amount of work as if they were hotspot users if the router is sending accounting information.

[cybermaus]

Yes, but there is no logon, no radius, no credit card stuff. There is a DHCP request, but thats Coova internal/local. It seems Coova does report the result, but there is no real load on your services. Hence my suggestion to count transactions rather than sessions. Especially since even 'free' logon cause a transaction.

And the way I understand it now, even if I did not have a macauth setup (and its really just the AP's themselves plus proprietors own phone), its counting every time someone gets out and back in range, or every time his/her Android phone goes to sleep and wakes up and re-requests DHCP, etc. This will surely cause the session counter to go up way too fast. And I am arguing this while already having subscribed to the upgrade, so I am giving honest advise, not just trying to get something for free.

[Roland]

No, you are counting it wrong. Yes, transaction and sessions are two different things. Transaction is when a user gets a package (whether its free or paid). Session is when a user logs in successfully. Every time a user logs in with a correct username and password, a new session is created. The amount of downloaded and uploaded megabytes are recorded and also the logged in time. It seems that coova is sending these to our servers even though the user is counted as 'local' user. When someone is just "catching your signal" and connects to your signal, that doesn't count. What counted is when a user logs in with a valid username and password and a new session is sent to our servers. Coova is sending local user's sessions to our servers incorrectly as local users shouldn't be reported.

"there is no real load on your services" - that's not true as our servers are receiving and recording sessions with real time usage updates.

The reason is why we are counting sessions rather than transactions is that sessions are the real measurement of using our resources.

Let's see an example: someone is giving out monthly accesses for residents (lets say 100 monthly accesses per month). These people use the internet regularly, and logging in multiple times a day. Our system handles their logins, and their usage (and usage is reported to our servers in every minute, not just when the user logs out).  These people can create hundreds of sessions every day and our servers have to work heavily.

An other operator gives out 30 minutes of accesses, 100 per month. Users usually use up 30 minutes with only 1 session.

Both are 100 transactions, but the number of sessions and usage of our resources are very different, in the first case it can be multiple thousands, in the second case it can be as less as 100. What a difference.


And to be honest, usage of local users should be counted if we want to give you usage statistics. Otherwise you should see false statistics about sessions and usage.

In my opinion you should change the free package to a bigger package, like 24 hours as your current setting (1 hour) kicks out users after 1 hour and generate a new session in every hour.

[cybermaus]

I sort of liked the 1 hour idea, because it would prevent users from leaving their laptops running inside the mobile homes/cottages while they are out, prevent smartphones from starting stuff while in their pockets, generally remind guests that WiFi is free but limited, and most of all, would prevent kids from trying to torrent movies.

I will leave it on one hour session time for now, and see how the season goes. I'll increase the session time if/as needed.