Re: [Dime] WGLC #1 for draft-ietf-dime-load-02

["A. Jean Mahoney" <[e-mail]>]

Hi Maria,

OK, I see your point. I keep thinking that since DNS is so cool everyone 
would use it...

On 7/18/16 2:25 AM, Maria Cruz Bartolome wrote:
> MCRUZ 2> The issue is to be able to provide a LOAD value that allows
> the client to perform load distribution. If we do not take the weight
> into account, somehow (implementation dependent), the distribution
> will be very far from even, it may cause very important traffic
> oscillations (e.g. small servers will appear as low loaded but if
> traffic is sent towards then they may reach overload threshold very
> soon) and big server will normally be underutilized. Therefore, the
> expected load distribution is far from being achieved.

In section 5, the implementer is told to find server capacity via the 
DNS SRV query and then use it to calculate load:

    The goal is make it possible to use both the load values received as
    a part of the Diameter Load mechanism and weight values received as a
    result of a DNS SRV query.  As a result, the Diameter load value has
    a range of 0-65535.  This value and DNS SRV weight values are then
    used in a distribution algorithm similar to that specified in
    [RFC2782].

But what if the Diameter node does NOT use DNS [1]? According to RFC 
6733, Section 5.2, DNS is "MUST be supported" but only "MAY be used". So 
where does the peer get the weight information if all of its connections 
are manually configured? I guess weights could be manually configured, 
but that becomes difficult to maintain as peers are upgraded.

So the question becomes:  Do we make use of DNS a MUST for this 
solution, or do we have reporting peers include capacity in their load 
reports?

Thanks!

Jean

[1] I've seen large-scale deployments with only manual configuration for 
all Diameter connections, with no DNS servers in the network.