| 花了一天时间，将原先放在免费虚拟主机上的 Flex 留言板，连同原先的数据库数据，全部搬家到了 Google App Engine 上，这里大概说说体会： 1、选 Python 还是选 Java ? |
2、GAE 速度怎么样？
说实话，很一般，Google 的粉丝不要拍砖哦，我这有证据。下面的两个网址，是同一个的 Flex 留言板程序，数据也一样。
第一个在 GAE 上
第二个在美国的一个免费虚拟主机上，距离远，而且速度带宽都有限制，自己比较下吧，我这里GAE 要慢很多。
要是同国内付费的虚拟主机比较，可能差距更大。
Google App Engine
普通虚拟主机
3、GAE 数据库好使吗？
还不错。但是 Google 应该在现有的 API 上再简单的包装一层，立刻就可以减少 80% 的重复代码。
比如，原先 PHP 版的留言板，在用 CodeIgniter 框架的情况下，一条留言从 request 到插入数据库，只需要几条语句。而现在要几十条语句。只要简单再包装一下，应该可以达到类似的效果。
4、GAE 配额都有什么限制？
限制的种类非常多，但是配额数都够用了。经过三四天的折腾，CPU 也才用了0.06个小时。
其他包括各种 API 调用次数也都是有限制的，我为了搬家，抓取原来的数据，使用了16次 UrlFetch API，好在配额六十多万次，够用了。
程序上传发布一天只能250次，我现在用了9次，问题也不大。
用的最多的是 Datastore API ，已经用了一千多次了，但是配额是每天一千万次，哈哈，够用了吧。
Requests
Quotas are reset every 24 hours. Next reset: 5 hours
| Resource | Daily Quota | Rate |
||
|---|---|---|---|---|
| CPU Time |
|
0% | 0.11 of 46.30 CPU hours | Okay |
| Requests |
|
0% | 2142 of 1333328 | Okay |
| Outgoing Bandwidth |
|
1% | 0.07 of 10.00 GBytes | Okay |
| Incoming Bandwidth |
|
0% | 0.00 of 10.00 GBytes | Okay |
| Secure Requests |
|
0% | 0 of 1333328 | Okay |
| Secure Outgoing Bandwidth |
|
0% | 0.00 of 10.00 GBytes | Okay |
| Secure Incoming Bandwidth |
|
0% | 0.00 of 10.00 GBytes | Okay |
Datastore
| Datastore API Calls |
|
0% | 4024 of 10368000 | Okay |
| Stored Data |
|
0% | 0.00 of 1.00 GBytes | Okay |
| Data Sent to API |
|
0% | 0.00 of 12.00 GBytes | Okay |
| Data Received from API |
|
0% | 0.02 of 116.00 GBytes | Okay |
| Datastore CPU Time |
|
0% | 0.09 of 62.11 CPU hours | Okay |
| Mail API Calls |
|
0% | 0 of 7000 | Okay |
| Recipients Emailed |
|
0% | 0 of 2000 | Okay |
| Admins Emailed |
|
0% | 0 of 5000 | Okay |
| Message Body Data Sent |
|
0% | 0.00 of 0.06 GBytes | Okay |
| Attachments Sent |
|
0% | 0 of 2000 | Okay |
| Attachment Data Sent |
|
0% | 0.00 of 0.10 GBytes | Okay |
UrlFetch
| UrlFetch API Calls |
|
0% | 0 of 657084 | Okay |
| UrlFetch Data Sent |
|
0% | 0.00 of 4.00 GBytes | Okay |
| UrlFetch Data Received |
|
0% | 0.00 of 4.00 GBytes | Okay |
Image Manipulation
| Image Manipulation API Calls |
|
0% | 0 of 864000 | Okay |
| Data Sent to API |
|
0% | 0.00 of 1.00 GBytes | Okay |
| Data Received from API |
|
0% | 0.00 of 5.00 GBytes | Okay |
| Transformations executed |
|
0% | 0 of 2592000 | Okay |
Memcache
| Memcache API Calls |
|
0% | 0 of 8640000 | Okay |
| Data Sent to API |
|
0% | 0.00 of 10.00 GBytes | Okay |
| Data Received from API |
|
0% | 0.00 of 50.00 GBytes | Okay |
Deployments
| Deployments |
|
0% | 0 of 250 | Okay |
最后看看超出配额的价格，下面的表格是每天0.1美元预算、偏重于存储的情况，11G的空间，够用了，博客园干脆也搬到 App Engine 上吧，呵呵。
| Resource (% of Budget) | Budget | Unit Cost | Paid Quota | Free Quota | Total Daily Quota |
|---|---|---|---|---|---|
| CPU Time 40% | $0.05 | $0.10 | 0.50 | 46.30 | 46.80 CPU hours |
| Bandwidth Out 8% | $0.00 | $0.12 | 0.00 | 10.00 | 10.00 GBytes |
| Bandwidth In 2% | $0.00 | $0.10 | 0.00 | 10.00 | 10.00 GBytes |
| Stored Data 50% | $0.05 | $0.005 | 10.00 | 1.00 | 11.00 GBytes* |
| Recipients Emailed 0% | $0.00 | $0.0001 | 0.00 | 2000.00 | 2000.00 Emails |
欢迎大家来我的新家疯狂留言，进行残酷的压力测试，看看能不能把它刷爆，呵呵。
我的GAE主页是：http://myqiao-app.appspot.com/exercise/
==================================================
本文转自左洸博客园博客，原文链接：http://www.cnblogs.com/myqiao/archive/2009/05/12/1454670.html，如需转载请自行联系原作者
