HydraGPS load balancing was RE: [load balancing] Verisign and Load Balancers

From: Nimesh Vakharia (nvakhariIZZATgenx.net)
Date: Sat Feb 24 2001 - 02:32:16 EST

  • Next message: Clement Lyons: "(no subject)"

    This is an interesting implementation but if I understand
    correctly u'r are not optimizing delivery from the client to the
            Correct me if I'm wrong but the HydraGPS's are independently
    located and the AS thats announcing the VIP (VIP1) is discontiguous.
    The GPS caluates the metric b/w itself and the probe/site

    </quote> "{{measured latency (delay) proximity (hop count and MTU),
    congestion (packets per second, lost packets, out of sequence packets),
    anomalies (duplicate and damaged packets), ICMP errors}}" </quote>

    rewrites the client requests and directs it towards the optimal
    destination (destination IP is the managment network VIP ie VIP2). The
    local SLB solution balances and responds back with the actual VIP1 (VIP1
    is not routable to this farm from the net).

    Feel free to correct this setup if you find any inconsistencies.

    Here are the problems as I see them:

    - the return path of the packets (the bulk of the traffic) ie from the
    server farm to the cleint has nothing to do with the metrics being
    measured between the GPS and the probe. For all you know the return path
    goes through a bunch of completely different service providers. You could
    be acheiving minimal to zero or even performance degradation at this

    - ISP tend to filter ingress traffic to prevent spoofing. This filters are
    setup querying the RADB (similar concept as to filtering your BGP
    routes). This solution indicates that you are spoofing packets from
    the server farms. Getting around it ie falsely modifying the RADB would
    break a lot of things on the net...

    Can you address these issues? I looked around your site but could not find
    any information on these cases.

    I used the FAQ's/PDF's as my source of information.

    explains the network design/layout for people who are interested.


     On Fri, 23 Feb 2001, G. Jules Huang wrote:

    > HydraWEB's new product HydraGPS avoids those DSN pitfalls you
    > mentioned.
    > It's currently in Beta Trial. For more information go to our
    > web site: http://www.hydraweb.com/products/hydragps/index.html
    > Jules
    > -----Original Message-----
    > From: owner-lb-lIZZATvegan.net [mailto:owner-lb-lIZZATvegan.net]On Behalf Of KJ
    > & JC Salchow
    > Sent: Wednesday, February 21, 2001 5:28 PM
    > To: lb-lIZZATvegan.net
    > Subject: Re: [load balancing] Verisign and Load Balancers
    > Speaking of Microsoft (and I'll probably have to turn in my MCSE lapel pin
    > after this), has anyone else been affected by the seemingly weird
    > conventions used in IE 5.0 in relation to Load Balancing? Of course after
    > all of the SSL talk - which from an LB standpoint came from the short TTL on
    > the SSL ID cache implemented in IE 5.0 - I'm sure we have all been bitten,
    > but I just found out something new that I'd like some ideas on.
    > This is about Global Load Balancing - or load balancing between multiple
    > sites (subnets?). I recently discovered that IE 5.0 (I haven't heard of
    > this before hand) caches DNS entries for 30 minutes by default. The bad
    > thing is that even if a site is down, it still hangs on to the address for -
    > 30 minutes. Many, if not all, of the global load balancing methods require
    > that during the initial connect - a DNS reply with multiple IP addresses is
    > supplied - that way if the primary is dead, the browser should use the
    > secondary address. This is basic DNS.
    > Microsoft's browser no longer does this. It takes one IP address and one
    > only. Additionally, it will not ask again until the cache on the first one
    > expires. I have recently seen a test case of this - Netscape worked
    > perfectly.
    > In addition, earlier today, while attempting to authenticate to a protected
    > site - an IE browser prompted for my authentication, but after it was
    > denied, would not prompt me again until I closed the browser and opened a
    > new one.
    > I am sure that all of these items can be corrected by changing settings in
    > the browser (registry), but if these are the default settings, things are
    > going to get ugly. Anyone else seen these things or have some global
    > solution?
    > Thanks
    > Ken

    This archive was generated by hypermail 2b30 : Sat Feb 24 2001 - 02:24:06 EST