CMU 推出 Product Management 的課程

CMUCS (Computer Science) 發的新聞稿:「Carnegie Mellon Offers New Master's Degree in Product Management」。

副標也清楚寫出是一年的課程:

One-Year Program Turns Computer Professionals Into "CEOs of the Product"

除了 CMU CS 外,也結合了 CMU 的 Tepper Business School 一起開:

A joint program of the university's School of Computer Science (SCS) and Tepper School of Business, the Master of Science in Product Management (MSPM) program will start January 2018.

另外一個不同角度的 Product Management。

CodeDeploy 對發生衝突的調整

AWS CodeDeploy 推出了新的功能,當 CodeDeploy 偵測到發生衝突時可以設定處理的方式:「AWS CodeDeploy Adds File Handling Support」。

以往遇到衝突時就會失敗,而現在可以選擇處理方式了:

Previously, CodeDeploy would fail the deployment when it detected a discrepancy between the files present at a target location and those in the last successful deployment. Now, you can choose how CodeDeploy responds (e.g., fail the deployment, retain the content, or overwrite the content) when encountering such files for each deployment.

這樣方便不少啊,之前在測試階段時偷改東西常常被抓包... XD

Amazon EC2 要推出 x1e.32xlarge,4TB RAM 的機器

剛剛 Amazon EC2 公佈的消息,要再推出記憶體大怪物機器 x1e.32xlarge:「EC2 In-Memory Processing Update: Instances with 4 to 16 TB of Memory + Scale-Out SAP HANA to 34 TB」。

Later this year we plan to make the x1e.32xlarge instances available in several AWS regions, in both On-Demand and Reserved Instance form. These instances will offer 4 TB of DDR4 memory (twice as much as the x1.32xlarge), 128 vCPUs (four 2.3 GHz Intel® Xeon® E7 8880 v3 processors), high memory bandwidth, and large L3 caches.

雖然都是拿 SAP 來舉例,但這也是當系統碰到瓶頸時搶時間的方法 (也就是當 happy problem 發生時的 workaround)。

不知道這種機器實際買起來多少錢 XDDD

美國聯邦法院認定 Open Source License 是有效的強制性合約

這是是韓國廠商 Hancom 在加州被告上法院:「A federal court has ruled that an open-source license is an enforceable contract」。

But after it began using Ghostscript in its software in 2013, Hancom did neither: it did not open-source its software, and it did not pay Artifex a licensing fee.

At the end of 2016, Artifex filed a lawsuit against Hancom in the US District Court for the Northern District of California.

Hancom 的理由是授權 (license) 不是合約 (contract):

That happened when Hancom issued a motion to dismiss the case on the grounds that the company didn’t sign anything, so the license wasn’t a real contract.

而法官認定 GPL 是合約的一種:

“Not so,” said Judge Jacqueline Scott Corley in her order on the motion on April 25. Corley said the GNU GPL “provides that the Ghostscript user agrees to its terms if the user does not obtain a commercial license. Plaintiff alleges that Defendant used Ghostscript, did not obtain a commercial license, and represented publicly that its use of Ghostscript was licensed under the [GNU GPL]. These allegations sufficiently plead the existence of a contract.”

HP 的 audio driver 內含 Keylogger

HP 被發現 2015 年簽出來的 audio driver 內含 keylogger (yeah,因為有數位簽名,所以賴不掉...):「[EN] Keylogger in Hewlett-Packard Audio Driver」。完整的報告在「modzero Security Advisory: Unintended/Covert Storage Channel for sensitive data in Conexant HD Audio Driver Package. [MZ-17-01]」這邊。

keylogger 記錄後會寫到 local file 裡。來拉板凳...

InnoDB redo log 大小對效能的影響

在「Benchmark(et)ing with InnoDB redo log size」這邊看到在討論 InnoDB redo log 的大小對效能的影響 (也就是 innodb_log_file_sizeinnodb_log_files_in_group)。

開頭就有先提到重點,在新版 MySQL 裡,幾乎所有的情況比較大的 redo log 有比較好的效能 (平均值):

tl;dr - conclusions specific to my test

  1. A larger redo log improves throughput
  2. A larger redo log helps more with slower storage than with faster storage because page writeback is more of a bottleneck with slower storage and a larger redo log reduces writeback.
  3. A larger redo log can help more when the working set is cached because there are no stalls from storage reads and storage writes are more likely to be a bottleneck.
  4. InnoDB in MySQL 5.7.17 is much faster than 5.6.35 in all cases except IO-bound + fast SSD

可以看出來平均效能的提昇很顯著,不管是增加 redo log 大小還是升級到 5.7:

但作者也遇到了奇怪的效能問題。雖然平均效能提昇得很顯著,但隨著加入資料的增加,效能的 degradation 其實很嚴重,在原來的網頁上可以看到這些資訊。

The results above show average throughput and that hides a lot of interesting behavior. We expect throughput over time to not suffer from variance -- for both InnoDB and for MyRocks. For many of the results below there is a lot of variance (jitter).

所以也許現階段先加大就好 (至少寫入的效能會提昇),不需要把這個特性當作升級 MySQL 的理由。

Linux 上跑電腦圍棋程式 (CGOS)

這邊講的不是對人下的,而是電腦之間的對弈。

目前大多數的對弈軟體會到「CGOS - A Go Server Just for Computers」這邊對弈,當然目前最紅的是一般 (人類) 正式比賽用的 19x19 棋盤,數子採用中國規則 (相當於貼 7.5 目)。

不過與常見的正式比賽比較不一樣的是採用包干制,每方限時 15 分鐘,超時就直接裁定敗,不過有個小例外:

CGOS silently adds a fraction of a second to each players clock for each move played.

要把圍棋程式接到 CGOS 上面需要兩段程式,一段是 Go Engine 本身要支援 Go Text Procotol (GTP),另外一段是把 GTP 接到 CGOS。

前面 Go Engine 的部份,目前不少圍棋軟體都有支援 GTP,像是 Leela 或是 Ray

後者一般會用 Python CGOS Client

其中比較特別的是 CGOS 的帳號密碼,帳號只允許 18 個字,另外沒有帳號申請系統,第一次用什麼帳號他就自動記錄起來,之後就要用這組。

# config.cfg
Common:
  KillFile = kill.txt

# First engine
GTPEngine:
  Name = Leela090-test
  CommandLine = ./leela090-gtp.sh

  ServerHost = yss-aya.com
  ServerPort = 6819
  ServerUser = Leela090-test
  ServerPassword = mypassword

  NumberOfGames = 1

  SGFDirectory = sgf

然後建立 sgf 目錄存棋譜後,用 python bin/cgosclient.py config.cfg 跑起來。當你 touch kill.txt 後,下一盤棋就會自己結束 (預設是一直下)。

再來是裡面提到的 leela090-gtp.sh

#!/bin/bash
exec /usr/bin/nice -n 20 $(dirname $0)/leela_090_linux_x64_opencl -g -t 8 -b 50 -q -l /tmp/leela090-gtp.log

這邊跑的是 OpenCL 版本,你也可以跑 CPU 版。其中 -g 是 GTP mode,-t 是 CPU thread 數量,-b 是 network latency penalty (避免超時),-q 是 quiet mode,-l 是 log。

MongoDB 的 consistent backup

PerconaGitHub 上放出 MongoDB 的 consistent backup 工具:「Percona-Lab/mongodb_consistent_backup: 1.0 Release Explained」。

程式碼在「Percona-Lab/mongodb_consistent_backup」這邊,首頁也講了對應的條件。

Percona 在 MongoDB 上投入愈來愈多資源了... (但我還是沒很想用 XD)

Backblaze 2017Q1 對硬碟的分析

Backblaze 放出 2017Q1 對硬碟的分析資料:「Hard Drive Stats for Q1 2017」。

相較於之前的報告 (Backblaze Hard Drive Stats for 2016),這次則是把硬碟數量考慮進去,做了一份有正負誤差的:

最近的趨勢沒什麼變,整體上來看 HGST 的品質還是最好的。

IE 與 Edge 推出更新,阻擋 SHA-1 憑證

微軟這幾天推出更新,IEEdge 將不會接受 SHA-1 憑證:「Microsoft Makes it Official, Cuts off SHA-1 Support in IE, Edge」。微軟的公告則是在「Deprecation of SHA-1 for SSL/TLS Certificates in Microsoft Edge and Internet Explorer 11」這邊。

根憑證不受影響 (所以企業自己產生的 Root CA 也不受影響):

Beginning May 9, 2017, Microsoft released updates to Microsoft Edge and Internet Explorer 11 to block sites that are protected with a SHA-1 certificate from loading and to display an invalid certificate warning. This change will only impact SHA-1 certificates that chain to a root in the Microsoft Trusted Root Program where the end-entity certificate or the issuing intermediate uses SHA-1. Enterprise or self-signed SHA-1 certificates will not be impacted, although we recommend that all customers quickly migrate to SHA-2 based certificates.