Re: Global snapshots

From: Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com>
To: "tsunakawa(dot)takay(at)fujitsu(dot)com" <tsunakawa(dot)takay(at)fujitsu(dot)com>, PostgreSQL-Dev <pgsql-hackers(at)postgresql(dot)org>
Cc: Masahiko Sawada <masahiko(dot)sawada(at)2ndquadrant(dot)com>, Amit Kapila <amit(dot)kapila16(at)gmail(dot)com>, "Andrey V(dot) Lepikhov" <a(dot)lepikhov(at)postgrespro(dot)ru>, "movead(dot)li(at)highgo(dot)ca" <movead(dot)li(at)highgo(dot)ca>
Subject: Re: Global snapshots
Date: 2021-01-08 12:21:53
Message-ID: 36d657e4-74aa-4f9c-3d6e-30c8a4a5a55f@oss.nttdata.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 2021/01/01 12:14, tsunakawa(dot)takay(at)fujitsu(dot)com wrote:
> Hello,
>
>
> Fujii-san and I discussed how to move the scale-out development forward. We are both worried that Clock-SI is (highly?) likely to infringe the said Microsoft's patent. So we agreed we are going to investigate the Clock-SI and the patent, and if we have to conclude that we cannot embrace Clock-SI, we will explore other possibilities.

Yes.

>
> IMO, it seems that Clock-SI overlaps with the patent and we can't use it. First, looking back how to interpret the patent document, patent "claims" are what we should pay our greatest attention. According to the following citation from the IP guide by Software Freedom Law Center (SFLC) [1], software infringes a patent if it implements everything of any claim, not all claims.
>
>
> --------------------------------------------------
> 4.2 Patent Infringement
> To prove that you5 infringe a patent, the patent holder must show that you make, use, offer to sell, or sell the invention as it is defined in at least one claim of the patent.
>
> For software to infringe a patent, the software essentially must implement everything recited in one of the patent�fs claims. It is crucial to recognize that infringement is based directly on the claims of the patent, and not on what is stated or described in other parts of the patent document.
> --------------------------------------------------
>
>
> And, Clock-SI implements at least claims 11 and 20 cited below. It doesn't matter whether Clock-SI uses a physical clock or logical one.

Thanks for sharing the result of your investigation!

Regarding at least claim 11, I reached the same conclusion. As far as
I understand correctly, Clock-SI actually does the method described
at the claim 11 when determing the commit time and doing the commit
on each node.

I don't intend to offend Clock-SI and any activities based on that. OTOH,
I'm now wondering if it's worth considering another approach for global
transaction support, while I'm still interested in Clock-SI technically.

Regards,

--
Fujii Masao
Advanced Computing Technology Center
Research and Development Headquarters
NTT DATA CORPORATION

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Simon Riggs 2021-01-08 12:33:59 Re: WIP: System Versioned Temporal Table
Previous Message Merlin Moncure 2021-01-08 11:52:20 Re: plpgsql variable assignment with union is broken