老刘测评
近期，UCloud优刻得“全球云产品，全场1折起UCloud全球33个数据中心，29条专线，覆盖五大洲，助力企业出海”海外云主机促销活动页面上香港云主机机型，统一由原先的通用型换成了快杰型，而其他海外可用区依然促销的是通用型云主机。
从UCloud官网控制台也可以看到同等配置快杰型要比通用型便宜一些，那么香港通用型和快杰型云主机性能差别大吗？带着这个疑问，老刘测评测试了促销页面上活动价格首年300元3年800元的1核2G内存2M带宽40G RSSD系统盘香港快杰云主机并与同等配置的香港快杰型云主机作了简单测评对比。下图是两种机型的机器基础信息。
测试工具：站长工具Ping值测试，SuperBench脚本测试和UnixBench脚本跑分。
Ping是潜水艇人员的专用术语，表示回应的声纳脉冲，在网络中Ping是一个十分好用的TCP/IP工具。它主要的功能是用来检测网络的连通情况和分析网络速度。
SuperBench脚本，可以一键测试云主机服务器VPS 各种性能，包括机器基本信息，硬盘读写速度，到国内各个地区的下载速度等。
UnixBench是一个类unix系（Unix，BSD，Linux）统下的性能测试工具，一个开源工具，被广泛用与测试linux系统主机的性能。Unixbench的主要测试项目有：系统调用、读写、进程、图形化测试、2D、3D、管道、运算、C库等系统基准性能提供测试数据，UnixBench运行结束后会给出机器整体性能跑分值，分数越高机器性能越好，一般跑分在1200-1300以上的机器，性能不错，值得入手。
Ping值测试对比
老刘使用站长工具Ping测试分别对通用型和快杰型云主机进行了测试以下是测试结果图：
通用型云主机Ping值
快杰型云主机Ping值
分析：由以上两图可以看出，通用型和快杰型云主机在电信、多线、联通和移动线路上响应时间基本一致，其中通用型平均49.2ms，快杰型平均48ms。
SuperBench测速对比
通用型云主机SuperBench测试结果
----------------------------------------------------------------------
Superbench.sh -- https://www.oldking.net/350.html
Mode : Standard Version : 1.1.7
Usage : wget -qO- sb.oldking.net | bash
----------------------------------------------------------------------
CPU Model : Intel Core Processor (Broadwell, no TSX)
CPU Cores : 1 Cores 2194.916 MHz x86_64
CPU Cache : 16384 KB
OS : CentOS 7.6.1810 (64 Bit) KVM
Kernel : 4.19.0-6.el7.ucloud.x86_64
Total Space : 1.5 GB / 40.0 GB
Total RAM : 98 MB / 1830 MB (429 MB Buff)
Total SWAP : 0 MB / 0 MB
Uptime : 0 days 7 hour 46 min
Load Average : 0.00, 0.00, 0.00
TCP CC : cubic
ASN & ISP : AS135377, UCloud (HK) Holdings Group Limited
Organization : UCLOUD INFORMATION TECHNOLOGY (HK) LIMITED
Location : Cheung Sha Wan, Hong Kong / HK
Region : Sham Shui Po
----------------------------------------------------------------------
I/O Speed( 1.0GB ) : 94.2 MB/s
I/O Speed( 1.0GB ) : 94.6 MB/s
I/O Speed( 1.0GB ) : 94.6 MB/s
Average I/O Speed : 94.5 MB/s
----------------------------------------------------------------------
Node Name Upload Speed Download Speed Latency
Speedtest.net 1.97 Mbit/s 48.75 Mbit/s 1.77 ms
Fast.com 0.00 Mbit/s 0 Mbit/s -
Nanjing 5G CT 1.94 Mbit/s 49.04 Mbit/s 34.48 ms
Hefei 5G CT 2.17 Mbit/s 48.94 Mbit/s 35.88 ms
Guangzhou 5G CT 1.98 Mbit/s 50.84 Mbit/s 11.26 ms
TianJin 5G CU 2.00 Mbit/s 60.17 Mbit/s 87.08 ms
Shanghai 5G CU 2.01 Mbit/s 59.27 Mbit/s 82.82 ms
Guangzhou 5G CU 2.01 Mbit/s 50.50 Mbit/s 6.12 ms
Tianjin 5G CM 1.91 Mbit/s 57.45 Mbit/s 86.61 ms
Wuxi 5G CM 2.01 Mbit/s 58.73 Mbit/s 43.50 ms
Nanjing 5G CM 2.10 Mbit/s 45.32 Mbit/s 47.25 ms
Hefei 5G CM 1.87 Mbit/s 59.18 Mbit/s 61.92 ms
Changsha 5G CM 2.15 Mbit/s 51.82 Mbit/s 30.56 ms
----------------------------------------------------------------------
Finished in : 6 min 28 sec
Timestamp : 2020-07-26 10:02:10 GMT+8
Results : ./superbench.log
----------------------------------------------------------------------
Share result:
· https://www.speedtest.net/result/c/1e22cbf6-b71e-43f8-a4df-bca5c1ea3d1a
· https://paste.ubuntu.com/p/kPpCH3WK4z/
----------------------------------------------------------------------
[root@10-8-74-26 ~]#
快杰型云主机SuperBench测试结果
----------------------------------------------------------------------
Superbench.sh -- https://www.oldking.net/350.html
Mode : Standard Version : 1.1.7
Usage : wget -qO- sb.oldking.net | bash
----------------------------------------------------------------------
CPU Model : Intel Xeon Processor (Cascadelake)
CPU Cores : 1 Cores 2294.608 MHz x86_64
CPU Cache : 16384 KB
OS : CentOS 7.6.1810 (64 Bit) KVM
Kernel : 4.19.0-6.el7.ucloud.x86_64
Total Space : 1.4 GB / 40.0 GB
Total RAM : 101 MB / 1830 MB (397 MB Buff)
Total SWAP : 0 MB / 0 MB
Uptime : 0 days 8 hour 55 min
Load Average : 0.00, 0.00, 0.00
TCP CC : cubic
ASN & ISP : AS135377, UCloud (HK) Holdings Group Limited
Organization : UCLOUD INFORMATION TECHNOLOGY (HK) LIMITED
Location : Cheung Sha Wan, Hong Kong / HK
Region : Sham Shui Po
----------------------------------------------------------------------
I/O Speed( 1.0GB ) : 77.7 MB/s
I/O Speed( 1.0GB ) : 78.1 MB/s
I/O Speed( 1.0GB ) : 78.1 MB/s
Average I/O Speed : 78.0 MB/s
----------------------------------------------------------------------
Node Name Upload Speed Download Speed Latency
Speedtest.net 1.98 Mbit/s 48.58 Mbit/s 1.67 ms
Fast.com 0.00 Mbit/s 48.3 Mbit/s -
Nanjing 5G CT 2.10 Mbit/s 48.61 Mbit/s 33.11 ms
Hefei 5G CT 2.27 Mbit/s 48.54 Mbit/s 43.72 ms
Guangzhou 5G CT 2.00 Mbit/s 49.98 Mbit/s 9.64 ms
TianJin 5G CU 2.10 Mbit/s 58.06 Mbit/s 99.49 ms
Shanghai 5G CU 2.03 Mbit/s 49.40 Mbit/s 42.65 ms
Guangzhou 5G CU 1.93 Mbit/s 50.77 Mbit/s 9.13 ms
Tianjin 5G CM 1.96 Mbit/s 57.22 Mbit/s 79.01 ms
Wuxi 5G CM 2.20 Mbit/s 54.79 Mbit/s 46.69 ms
Nanjing 5G CM 2.10 Mbit/s 47.14 Mbit/s 45.33 ms
Hefei 5G CM 1.89 Mbit/s 58.52 Mbit/s 62.17 ms
Changsha 5G CM 2.00 Mbit/s 52.15 Mbit/s 27.26 ms
----------------------------------------------------------------------
Finished in : 6 min 52 sec
Timestamp : 2020-07-26 11:10:03 GMT+8
Results : ./superbench.log
----------------------------------------------------------------------
Share result:
· https://www.speedtest.net/result/c/814e639b-5a69-44de-8304-89e71c825598
· https://paste.ubuntu.com/p/TGCXjmWmht/
----------------------------------------------------------------------
[root@10-8-109-48 ~]#
分析：在基本信息上可以看到通用型和快杰型云主机的CPU不同，另外Total Space和RAM略有区别，不过整体一致，机器配置如官方描述；通用型云主机的I/O读写速度在94.5MB/s，而快杰型在78.0MB/s，通用型在I/O磁盘读写性能方面更胜一筹；测速国内节点，通用型和快杰型基本一致，上行下行均能跑满，延时也无差别。
UnixBench跑分对比
通用型云主机UnixBench跑分结果
------------------------------------------------------------------------
Benchmark Run: 日 7月 26 2020 10:21:27 - 10:49:30
1 CPU in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 29195618.7 lps (10.0 s, 7 samples)
Double-Precision Whetstone 4425.1 MWIPS (9.9 s, 7 samples)
Execl Throughput 5106.8 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 1024533.7 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 270441.1 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 2809932.2 KBps (30.0 s, 2 samples)
Pipe Throughput 1640017.0 lps (10.0 s, 7 samples)
Pipe-based Context Switching 292713.6 lps (10.0 s, 7 samples)
Process Creation 14346.6 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 6114.9 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 818.8 lpm (60.0 s, 2 samples)
System Call Overhead 2063635.7 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 29195618.7 2501.8
Double-Precision Whetstone 55.0 4425.1 804.6
Execl Throughput 43.0 5106.8 1187.6
File Copy 1024 bufsize 2000 maxblocks 3960.0 1024533.7 2587.2
File Copy 256 bufsize 500 maxblocks 1655.0 270441.1 1634.1
File Copy 4096 bufsize 8000 maxblocks 5800.0 2809932.2 4844.7
Pipe Throughput 12440.0 1640017.0 1318.3
Pipe-based Context Switching 4000.0 292713.6 731.8
Process Creation 126.0 14346.6 1138.6
Shell Scripts (1 concurrent) 42.4 6114.9 1442.2
Shell Scripts (8 concurrent) 6.0 818.8 1364.7
System Call Overhead 15000.0 2063635.7 1375.8
========
System Benchmarks Index Score 1514.5
======= Script description and score comparison completed! =======
[root@10-8-74-26 ~]#
快杰型云主机UnixBench跑分结果
-----------------------------------------------------------------------
Benchmark Run: 日 7月 26 2020 11:32:43 - 12:00:46
1 CPU in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 39489607.9 lps (10.0 s, 7 samples)
Double-Precision Whetstone 4705.1 MWIPS (9.8 s, 7 samples)
Execl Throughput 5444.1 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 1150950.6 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 357391.6 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 2155191.0 KBps (30.0 s, 2 samples)
Pipe Throughput 2157118.7 lps (10.0 s, 7 samples)
Pipe-based Context Switching 356369.0 lps (10.0 s, 7 samples)
Process Creation 14155.1 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 7039.5 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 923.4 lpm (60.0 s, 2 samples)
System Call Overhead 2554603.7 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 39489607.9 3383.9
Double-Precision Whetstone 55.0 4705.1 855.5
Execl Throughput 43.0 5444.1 1266.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 1150950.6 2906.4
File Copy 256 bufsize 500 maxblocks 1655.0 357391.6 2159.5
File Copy 4096 bufsize 8000 maxblocks 5800.0 2155191.0 3715.8
Pipe Throughput 12440.0 2157118.7 1734.0
Pipe-based Context Switching 4000.0 356369.0 890.9
Process Creation 126.0 14155.1 1123.4
Shell Scripts (1 concurrent) 42.4 7039.5 1660.3
Shell Scripts (8 concurrent) 6.0 923.4 1539.0
System Call Overhead 15000.0 2554603.7 1703.1
========
System Benchmarks Index Score 1714.6
======= Script description and score comparison completed! =======
[root@10-8-109-48 ~]#
分析：UnixBench测试了系统调用、读写、进程、图形化测试、2D、3D、管道、运算、C库等性能，是一个比较综合性测评机器性能的工具，这里具体各个性能数据不用作分析，这里我们只需要看最后的跑分结果。通用型云主机跑分1514.5，快杰型云主机跑分1714.6，均是在1200-1300分值以上，性能优秀，同时快杰型云主机跑分更高，综合性能较通用型云主机好。
总结，老刘测评通过以上测试结果分析可知，UCloud优科得香港1C2G2M配置的通用型和快杰型在连通性和速度上基本无差别，通用型较快杰型在磁盘读写性能上更好些，而就云主机综合性能考虑，老刘建议选择快杰型云主机，其UnixBench跑分更高。
