[cs615asa] Meetup Summary

[Xiaoyun Liu]

Meetup: Multi-host, multi-network persistent containers
Link: https://www.meetup.com/DevOpsContained/events/249268845/
Date: Thursday, May 3, 2018
Speaker: Bruce Parker(Aerospike)

Dr. Parker is a senior solutions architect at Aerospike, Inc. And he gave
us a really great speech. The main topic of his speech is "Databases and
Containers: From Development to Deployment". Actually, before this Meetup,
I always confused about container, I thought it's really similar with
virtual machines and I didn't know exactly how it works. But after Dr.
Parker's explanation, I understand that containers is an union read-write
file system, all the layer is only readable except the first layer. And it
includes everything needed to run itself. Containers also available for
both Linux and Windows based apps, and the software will always run the
same regardless of the environment. Compare to virtual machines, containers
are more portable and efficient because containers virtualize the operating
system instead of hardware.

Then Dr. Parker talked about the technology they used in Aerospike.
Aerospike it naturally suited to container deployment because
-Shared nothing architecture
-Automatic hashing of keys across the cluster (using the RIPEMD-160
collision free algorithm) with Smart Partitions
-Automatic healing of the cluster as nodes enter and leave the cluster,
including automatic re-balancing
-Automatic discovery of the cluster topology with Aerospike Smart Client
for all the populate languages and frameworks
-Automatic replication of data across nodes with a customizable replication
factor
In this way, we could Scale (up and out) the persistence layer, eliminate
reconfiguration of the application and database tier as Containers enter
and leave the topology and utilize Containers on your own dedicated
infrastructure or public cloud providers.

They handle data with really large size and meet the requirement of backup,
dynamic self discovery and cluster formation, application tier also deal
with "black box". Which it's kind of difficult, and there're still some
problem we need to solve. For example, the comparison of Docker and
Virtualization. We discussed the clustering, include multiple DNS entries
and single DNS entry. He also introduced three storage method using
container:
-Inside containers encapsulation of concerns.
-Outside containers separate of concerns.
-Data containers managed like other containers, also have special rules for
destruction, but with unknown performance.

Dr. Parker said that containers plus databases makes happy developers,
ephemeral containers plus databased equals to DevOps headaches. And I think
he made a great point. This is a great meetup and I really have a great
time. The people there are so nice. And really interesting thing is that
Dr. Parker happened to be the teacher of Stevens several years ago, we had
great conversation with him. Even though I'm still new to this area, this
meetup really encourage me to learn more.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.stevens.edu/pipermail/cs615asa/attachments/20180504/72e2ad2a/attachment-0001.html>