Jump to content

Ticketday: Whats your Internetspeed? Importance?


tullux
 Share

Recommended Posts

  • Replies 190
  • Created
  • Last Reply

Top Posters In This Topic

They might have either multiple public IP's NAT'ed through to feed the load balancer itself then, or each public IP gets translated into a farm of n number of backend server clusters. We have just put in the same load balancer here as See were using in 2015 at least for the sale. Might have a look at it before ticket day to get a bearing on how it works..... ;)

Link to comment
Share on other sites

7 minutes ago, Pinhead said:

They might have either multiple public IP's NAT'ed through to feed the load balancer itself then, or each public IP gets translated into a farm of n number of backend server clusters. We have just put in the same load balancer here as See were using in 2015 at least for the sale. Might have a look at it before ticket day to get a bearing on how it works..... ;)

Riverbed Stingray?

I think the DNS records returned are the load-balancers themselves....advertised through DNS on a round-robin basis with a 60 second TTL.  There will then be several webheads behind each on private IP's running the booking app all being served by a backend database.

Link to comment
Share on other sites

14 minutes ago, Pinhead said:

Owned by Broacade now, but same product line yep.

Yep, sounds like the likely architecture. The low TTL will be so they can change them quickly if they need to as they will only be cached by DNS clients for 60 secs.

Is it still an IP belonging to Virgin's network?

 

The usual 194 addresses belong to Virgin.....but interestingly the 109 addresses that are currently being resolved to belong to another cloud hosting provider, exponential-e.

I'm guessing exponential-e may host the DR platform to avoid Virgin being a single point of failure.

Link to comment
Share on other sites

Yep - I reckon DR is a good guess for the second lot.

Thinking back to when this discussion last went on in some detail a couple years ago, I remember See were identified as using a reputation filter at the front end as well as a sophisticated session pool manager at the back end (with the load balancer between these). May be worth looking up that thread I guess. If the architecture is largely unchanged then, the public IP's may possibly be NAT'ed by the reputation filter itself (probably part of a firewall product).

Just had a look at the architecture of the Brocade product here. First suggestion is to consider doing CTRL-F5 with this if it does sit immediately behind the public IP's...

Disclaimer: whether any of these suggestions either helps of hinders your ticket purchasing experience is not the responsibility of the poster!

Link to comment
Share on other sites

2 hours ago, Pinhead said:

 

Odd that See would present a public IP for any of the backend servers (behind the load balancer) at all really. You'd think it would be a single public IP (and DNS record) for the load balancer, and all the back end servers would be on an internal and unrouted network.

They didn't.

That year they had two user-facing servers, but with an set-up that required both to be publicly-accessible for a sale to be able to be completed, but only one of them was. The hosts file hack made the 2nd server accessible, and that caused the system to work.

I'm not entirely sure what it was precisely with the set-up that needed both servers to work, but it might be (hopefully someone can confirm/deny?) something as simple as a DNS with two IP entries causes the user's traffic to alternative between them, meaning that for the next page you'll always require access to the other server.

(PS: that was the system as it was that year. it has a different frontend load-balancing set-up now).

Edited by eFestivals
Link to comment
Share on other sites

1 hour ago, eFestivals said:

They didn't.

That year they had two user-facing servers, but with an set-up that required both to be publicly-accessible for a sale to be able to be completed, but only one of them was. The hosts file hack made the 2nd server accessible, and that caused the system to work.

I'm not entirely sure what it was precisely with the set-up that needed both servers to work, but it might be (hopefully someone can confirm/deny?) something as simple as a DNS with two IP entries causes the user's traffic to alternative between them, meaning that for the next page you'll always require access to the other server.

(PS: that was the system as it was that year. it has a different frontend load-balancing set-up now).

Ah, I get it - this would have been the situation previous to adding a load balancer (with a public cluster IP for all the backend servers). Are you saying though that the the success of a transaction on one server was dependent on the other being available as well? I thought at the time the issue was where they had two identical and independent transaction processing servers, but only one could be reached. That one got overloaded and stopped responding altogether. Then the DNS was corrected (or people browsed straight to the IP rather than the FQDN), and the other unloaded server suddenly became available and loads of peeps were able to get tix really quickly because of that (until it became busy again). With DNS Round Robin there is no guaranteed alternation between the IP's listed to the same A record in the zone (there is no weighting that can be added to an A record to my knowledge), so its entirely random which one a DNS client OS will chose. The unlucky I guess would hit the working server, but then upon next DNS query the client could be redirected to the unavailable server and the process would fail. If you were lucky, you could keep hitting the working server throughout the process all the way to the end. I would have thought that dependent on the TTL to some extent, which if lower would cause more frequent DNS queries introducing a higher probability of being redirected to the unavailable server before the transaction had completed. Thankfully they have a much improved system now eh.

Link to comment
Share on other sites

Just now, Pinhead said:

Heh - I think I'm in nerd mode cos I'm at work and have to do nerdy stuff like this all day unfortunately....

It's like trying to have a conversation with my brother :lol:

Trouble is because I'm in IT procurement and work on data centre projects and the like, he assumes I know what he's on about.  I've got enough knowledge not o look like a complete imbecile, but at the end of the day I just buy the stuff!

Link to comment
Share on other sites

In my experience internet speed counts for nothing in the ticket sales. I say this as someone who has used 50MB and didn't get tickets, and a max of 2MB and got tickets. I've also got tickets using 4G. Fancy tech isn't a factor in getting tickets, it's very much if the page is free when you click F5. It's luck :)

I opt for IE, Chrome, Firefox open, F5 on Chrome and let the others 20 second countdown. I've found if I get through on one, all of them turn into a booking page. I know some people use an auto-refresh type app on Chrome. I opt for the madness of repeating F5 over and over until I'm told it's sold out or get through! I tend to come off social media and forums etc just to see how far I manage to get. Usually there's two of us in the same house trying on different computers. This time we are trying from the house and the US. So at least we will be on different internet providers and hopefully one of us will have luck. Is there an international phone line still?

Link to comment
Share on other sites

I managed to get through last year using my mums shitty Windows PC that has really slow internet so I have every confidence that I'll do the same this year **

It's vitally important that I get tickets this year. If 2018 is a fallow year then I'll miss 2 years in a row.

 

**actually that's a lie, I'm shitting myself as always.

Link to comment
Share on other sites

18 hours ago, Pear_Cider said:

There hasn't been a telephone booking line for a couple of years now sadly, either domestic or international. I say sadly because I managed to get through on the phones a couple of times in the past.

Nevermind in that case, will be two computers, different locations in the world. Never know, different F5 pressing rates might be helpful.

Link to comment
Share on other sites

On 9/20/2016 at 11:47 AM, parsonjack said:

Oh, and if you want to leave a window auto-refreshing do it on another device, or at least a different browser on same device - simply using a second tab in the same browser won't give you any advantage as it'll refresh every time you manually refresh the other.

Sandbox (isolate) the session by using incognito / private browsing windows! 

Not just useful for watching porn and hiding your history! :ph34r:

Link to comment
Share on other sites

54 minutes ago, tullux said:

i now own a hybrid router which will have LTE in addion, if the "normal" internetspeed will not be reached.

 even if the speed does not matter, it seems that LTE has more chances, so its the best of both worlds.

What does that do to your IP address?  Does it change during a session?  I'm intrigued.  Do you happen to know the model name of the router?

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Unfortunately, your content contains terms that we do not allow. Please edit your content to remove the highlighted words below.
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
 Share

  • Recently Browsing   0 members

    • No registered users viewing this page.



×
×
  • Create New...