Truely Hot Fail over Multiple wan link possible in balance 30?


#1

Hello. I watched the interesting video about cut off the cable during wan connection. it was so interesting and exactly was what i want. but i think it was introduction about speed-fusion vpn technology.
Also i heard that fail-over 100% works for outgoing traffic through peplink loadbalance router.
actually i want to know that if I buy balance 30, can I have truly hot fail-over about Internet over multiple wan connection (and does it maintain the current user session)?
Does it switch back automatically for users too or some user action require too? and how long the fail-over time does it takes for switching automatically for users?

let me explain it more what really i want. well i’m looking for truly hot fail over solution, means that if for one of link problem occurred another link switch automatically well. of course many router does this but actually does not maintain the current client’ session and connection.
let me take an example to make it more bright. consider you try to continues ping to one valid host (ping - t 4.2.2.4) from your PC (that connect to balance 30 pep-link router). two wan connected to your pep-link router. at one time one link interrupt suddenly, so what will happen for the PC user? Does the ping reply will show “request timed out” and will not switch back automatically OR it does automatically and you will not see the ping failed message? (the result can be also acceptable even the fail-over time can be small like some few ping request timed out show to user too)-> like less than 10 failed icmp reply.
I know in some router fail-over possible but user action required. in fact fail-over occurred correctly just about new connection not before. for example with above example you should interrupt your ping one time and again ping it. it will be successful! but not about the previous connection that you have already established before interruption.
This is most important to know if any user action needed and the current session was maintained by the router or not? I also show it bellow too. Thanks in advance.

Truly hot failover (what i mean) :

ping -t 4.2.2.4 (from pc user)
Reply from 4.2.2.4: bytes=32 time=92ms TTL=41
Reply from 4.2.2.4: bytes=32 time=92ms TTL=41
Reply from 4.2.2.4: bytes=32 time=161ms TTL= 50 \ <- Link Failure Happened & Switched Automatically
Reply from 4.2.2.4: bytes=32 time=161ms TTL=50
Reply from 4.2.2.4: bytes=32 time=161ms TTL=50

OR
Successful fail-over with small down -time (for swiching link)

ping -t 4.2.2.4 (from pc user)
Reply from 4.2.2.4: bytes=32 time=92ms TTL=41
Reply from 4.2.2.4: bytes=32 time=92ms TTL=41
Reply “Ping requets timed out” \ <- Link Failure Happened & Switched Automatically
Reply "Ping requets timed out"
Reply from 4.2.2.4: bytes=32 time=161ms TTL= 50
Reply from 4.2.2.4: bytes=32 time=161ms TTL=50
Reply from 4.2.2.4: bytes=32 time=161ms TTL=50

if you have any experience let me know which one will happen?
Thanks.
Ali


#2

Hot Failover is part of the SpeedFusion VPN technology.

Hot Failover requires you to use SpeedFusion VPN, there is no other way to enable true hot failover.
This means you will need make a SpeedFusion VPN tunnel to a location with a second Peplink device if you want 0 downtime.

The Peplink Balance 30 does not support Hot Failover.
Hot Failover is supported from the Peplink Balance 210 and up, or from Peplink Balance One with optional software license.
I’ve edited down the Peplink Balance comparison matrix for you, to make it relevant to this post:

Standard failover will always give you some downtime, since all traffic has to be redirected to the secondary WAN first.
With SpeedFusion Hot Failover, this is not the case, since the secondary WAN is also actively being used in the VPN tunnel.


#3

Dear Joey, Many thanks for your complete answer. you replied to many of my questions. just small question remains, i was wondering if help me to find out theme too. here it is:
I also agree with you, and think your talking about Speed fusion and other VPN related Technology and VPN is completely true, but in this case what will be happen for your Internet connection?
let me explain it more. ofcourse hot-failover for vpn is very important and is a great job. but actually each company usually use internet connection that ISPs also usually doesn’t offer vpn connection for accessing their Services. ** what I found now is that in such usual internet case (dhcp,ppoe, static ip) we shouldn NOT expect hot-failover. and we have down-time.
Ok. now there is two important questions:

  1. Does switching back to healthy Internet link take “automatically” in current user connection (like ping -t 4.2.2.4) or user-action required?
  2. how long the fail-over time does it take to switch to appropriate healthy internet link (how long is fail-over time does it take)?

Let me show what exactly i have mentioned in my quesion:


Senario 1) The router maintain existing user session, no user action needed and it will be switched automatically

.*
My just few question are here:

Client PC> ping 4.2.2.4 -t \\ User Ping from his PC in LAN that connected to Pep-link 30 Balance router
Client PC> Reply from 4.2.2.4: bytes=32 time=93ms TTL=49 \\ Ping Replied well
Client PC> Reply from 4.2.2.4: bytes=32 time=93ms TTL=49 \\ …
Client PC> Reply from 4.2.2.4: bytes=32 time=93ms TTL=49 \\ *One Internet link drop suddenly After that!
Client PC> Request timed out. \\ Ok, We are in fail-over time
Client PC> Request timed out. }
Client PC> Request timed out. |
… > … … \\ How long is fail-over time approximately ?
Client PC> Request timed out. |
Client PC> Request timed out. }
Client PC> Request timed out. \\ (almost nothing untill here for user)
Client PC> Reply from 4.2.2.4: bytes=32 time=*112 ms TTL=51 \\ Fail-over done by auto switching to next link
Client PC> Reply from 4.2.2.4: bytes=32 time=112 ms TTL=51
Client PC> Reply from 4.2.2.4: bytes=32 time=112 ms TTL=51 \\ the fault link become up now and all links up!
Client PC> Request timed out. }
Client PC> Request timed out. \\ switching back to healthy link takes place with down time again?
Client PC> Request timed out. }
Client PC> Reply from 4.2.2.4: bytes=32 time=93ms TTL=49 \\ The internet link recovered completely and to best choice switched auto.

**

Senario 2) The router Does NOT maintain existing user session, user action

**
REQUIRED but it also switched automatically*.
Its almost bright.

Client PC> ping 4.2.2.4 -t
Client PC> Reply from 4.2.2.4: bytes=32 time=93ms TTL=49 \ Ping Replied well
Client PC> Reply from 4.2.2.4: bytes=32 time=93ms TTL=49 \ …
Client PC> Reply from 4.2.2.4: bytes=32 time=93ms TTL=*49 \*One Internet link drop suddenly!
Client PC> Request timed out. \ …
Client PC> Request timed out. \ * User doesn’t any feel about uptime even the next healthy link is available
Client PC> Request timed out. \ Fault link was recovered and up now. Both Links are now available
Client PC> Request timed out. \ Nothing happens for the LAN user
Client PC> Request timed out. \ the user still can not feel exactly if really link is up (or down)!
Client PC> Ctrl + C \ User see nothing so try to start new session and connection.
Client PC> Request timed out.
Client PC> Reply from 4.2.2.4: bytes=32 time=93ms TTL=49 \ Finally he found it replied successful and continue the job

I just take this example to show that maintaining the current connection how can be important during fail over.
Just my last point is we consider all of the senario are based through "Load Balance AND Fail-over solution" (NOT Just fail-over).
Thanks in advance.
Ali