Labs presents research on Edge Containers as a Service: ACM EdgeSys’21 Workshop

[Curt_Hopkins]

Hewlett Packard Labs researchers from HPE’s Network and Distributed Systems Lab (NDSL) took part in the ACM EdgeSys 2021 Workshop to discuss their latest research on edge systems, analytics, and networking. The workshop co-located with Eurosys Conference was virtually organized on April 26 from Edinburgh, Scotland.

Lianjie Cao, Anu Mercian, Diman Zad Tootaghaj, Faraz Ahmed and Puneet Sharma, along with Vinay Saxena from Hewlett Packard Enterprise presented their research paper titled  "eCaaS: A Management Framework of Edge Container as a Service for Business Workload" at this prestigious workshop. This year’s workshop was very well attended with over 300 attendees.

“Lack of IT expertise at the Edge makes it absolutely essential for our customers that we provide zero-touch cloud-based deployment and management of full-stack, bare-metal to containerized applications,” says Vinay Saxena, CTO & Chief Architect of HPE GreenLake Cloud Services.

“Edge Container as-a-Service (eCaaS) is a research project we have been working on for a year,” says Cao.

eCaaS, according to the project’s executive summary “provides automated lifecycle management of containerized edge sites and applications. With eCaaS, users can create customized edge sites with only high-level business intents which are analyzed and translated to deployment templates with low-level specifications. The edge site deployment templates are then automatically executed to build, deploy, and configure the containerized edge sites and applications.”

According to Cao, the team initially presented the project at the HPE Discover 2020 and continued to work out details during the pandemic and wound up with a research tool that will allow a user deploying applications on edge servers – in say a factory, in retail, or in a restaurant – to manage those servers, including apps, without having to manually rebuild the entire stack every time.

“The workshop’s technical program committee, which reviews the submissions, appreciated the timeliness of our research as complex orchestration and ML tasks are now moving to edge locations,” says Sharma.

Labs’ research is timely because more and more edge deployments have requirements for complex orchestration which cannot be fulfilled by limited IT expertise available at the edge, according to Sharma. Some of this complexity comes from the fact that some of these edge deployments are for machine learning (ML) tasks, which are increasingly being deployed at the edge. eCaaS simplifies all this using automation.

“The coolest part of this project was solving such a hard problem and providing a holistic and converged management experience for customers who are not familiar with how to manage large-scale heterogeneous containerized edge sites,” says Zad Tootaghaj. “We transformed it from a difficult aspect of full-stack server management to a pain-free process.”

Automatically being able to bring up and manage an edge infrastructure without manual intervention by either an IT expert or by the user is a great part of what the eCaaS solution accomplishes, according to Ahmed. This “intent-based management” allows a user to specify intent in declarative fashion. “If you can express it, you can create it.”

Curt Hopkins
Hewlett Packard Enterprise

twitter.com/hpe_labs
linkedin.com/showcase/hewlett-packard-labs/
labs.hpe.com