HI，chairs Wu:
I will try to give a brief reply on this comments。
1. Can you provide details about network information and cloud information
exposure? you can provide parameters or attributes list for network information
and cloud information?
The network information exposure(e.g srv6-policy) : The keys of
srv6-policy and its abttibutes like delay, packet loss, jitter,bsid,etc。the
cloud informatios exposure: the cost of clould and Computational resources
like cpu and
memory,etc。
2. what schema description file do you use to collect network information
and cloud information? what format looks like?
Can you provide a JSON example for schema description file format?
The Json farmat looks likes as the following described：
3. How do you make sure the path can be managed in response to the scale
down or up of the cloud resource? How do you correlate network information with
cloud information in your case?
In the db-ors framewoks, there is a Comprehensive orchestrator of
cloud&&networks which will collect the informations exposed by the network
&&cloud in real time and then make a decison to control and adjust the
network paths。
4. Is the network resource is allocated dedicated for a specific cloud
service? what about network resource sharing by multiple cloud service?
The network physical resources are shared by mulitiple cloud
service,like link bandwidth ；The logical resources abstracted from network
physical resources are allocated dedicated for a specific
cloud service。
5. what ALTO protocol extension you are looking for to support cloud and
network integration in your cases such as Cloud Access, DCI? I am under
impression is that endpoint identified by IPv4/IPv6 address is not sufficient
to address your use cases, you might need to extend endpoint address?
Fine-grained abstractions of network physical resources are needed for
ALTO protocol extension，whether The distributed database is absorbed to the
alto wg can be discussed and evaluated in the future。 In the db-ors framewoks,
there is a definition named "link identification" which identifies the
abastred service,looks like as following:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Cloud-ID| R LT| Logic-id |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
drafts:
https://datatracker.ietf.org/doc/draft-zhou-alto-dbors-framework/
https://datatracker.ietf.org/doc/draft-zhou-alto-dbors-requirement-usecase/
周芬林
道阻且长，行则将至！
系统软件工程师
软件平台南京开发一部/有线研究院/系统产品_有线产品经营部
中兴通讯股份有限公司
南京研发中心中兴1期
T: 15861819442
E: [email protected]
www.zte.com.cn
原始邮件
发件人：QinWu
收件人：IETF ALTO;
抄送人：周芬林10312441;袁冬宇10335620;
日 期 ：2022年11月02日 14:33
主 题 ：Database-based Open Resource Service Framework, Use Case and Requirements
Hi, Authors of Database-based Open Resource Service drafts:
Thank for proposing your idea to ALTO WG, there are two relevant drafts
https://datatracker.ietf.org/doc/draft-zhou-alto-dbors-framework/
https://datatracker.ietf.org/doc/draft-zhou-alto-dbors-requirement-usecase/
These two drafts aim at network and cloud integration scenarios, I am more
interested in
how you implement data based open Resource service using ETCD, a few questions
and comments on these two drafts:
1. Can you provide details about network information and cloud information
exposure? you can provide parameters or attributes list for network information
and cloud information?
2. what schema description file do you use to collect network information and
cloud information? what format looks like?
Can you provide a JSON example for schema description file format?
3. How do you make sure the path can be managed in response to the scale down
or up of the cloud resource? How do you correlate network information with
cloud information in your case?
4. Is the network resource is allocated dedicated for a specific cloud service?
what about network resource sharing by multiple cloud service?
5. what ALTO protocol extension you are looking for to support cloud and
network integration in your cases such as Cloud Access, DCI? I am under
impression is that endpoint identified by IPv4/IPv6 address is not sufficient
to address your use cases, you might need to extend endpoint address?
-Qin_______________________________________________
alto mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/alto
