Earticle

현재 초등 인공지능 프로그래밍 교육에서 널리 사용되는 클라우드 기반 플랫폼들은 학생 데이터를 외부 서버로 전송하여 처리하는 경우가 많다. 이 과정에서 학생들의 웹캠 영상, 음성 녹음, 개인 사진 등 민감한 정보가 외부로 유출될 수 있는 개인정보 보호 문제가 발생한다. 본 논문에서는 이러한 문제를 해결하기 위해 교사가 직접 간단히 운영할 수 있는 웹소켓 서버의 기술적 타당성을 검증하였다. 2코어 프로세서와 4GB 메모리를 갖춘 저사양 클라우 드 환경에서 파이썬과 러스트로 구현한 웹소켓 서버를 테스트했다. 실험은 동시 접속자 수를 10명에서 300명까지 변경하며, 100바이트의 텍스트 데이터와 23킬로바이트의 이미지 데이터를 다양한 빈도로 전송하는 6가지 시나리 오로 구성되었다. 실험 결과, 일반적인 초등학교 학급 규모인 30명 환경에서 두 서버 모두 대부분의 시나리오에서 100밀리초 이하의 응답 시간을 달성하여 실시간 교육 활동에 충분한 성능을 보였다. 모든 시나리오에서 메시지 무결성과 순서 보장이 완벽하게 이루어졌다. 이러한 결과는 교사가 저렴한 클라우드 서버나 개인 노트북으로도 학 급 단위 실시간 협업 학습을 안정적으로 지원할 수 있음을 의미한다.

Currently, cloud-based platforms widely used in elementary AI programming education often transmit and process student data on external servers. This process raises privacy concerns as sensitive information such as students' webcam footage, voice recordings, and personal photos may be exposed to external entities. This paper validates the technical feasibility of WebSocket servers that teachers can easily operate themselves to address these issues. We tested WebSocket servers implemented in Python and Rust in a low-specification cloud environment with a 2-core processor and 4GB of memory. The experiments consisted of six scenarios with concurrent users ranging from 10 to 300, transmitting 100-byte text data and 23-kilobyte image data at various frequencies. The experimental results showed that in a typical elementary school classroom environment of 30 students, both servers achieved response times below 100 milliseconds in most scenarios, demonstrating sufficient performance for real-time educational activities. Message integrity and order preservation were perfectly maintained across all scenarios. These findings indicate that teachers can reliably support class-level real-time collaborative learning using affordable cloud servers or personal laptops.

Programming models for cloud computing has been a research focus recently. Current programming models for cloud computing mainly focus on improving the efficiency of the cloud computing platforms but little has been done on the performance predictability of models. This paper presents a cloud computing programming model, called BSPCloud. The BSPCloud has the advantages of predictable performance. BSPCloud uses a hybrid of distributed-memory and shared-memory bulk synchronous parallel (BSP) programming model. Computing tasks are first divided into a set of coarse granularity bulks which are computed by the distributed-memory BSP model, and each coarse granularity bulk is further divided into a set of bulk threads which are computed by the shared-memory BSP model. BSPCloud makes full use of the multi-core architecture, and more importantly, the performance of the BSPCloud is predictable. We have implemented a proof-of-concept BSPCloud parallel programming library in java. We give the implementation of BSPCloud parallel programming library and show how the library implements hybrid programming. We detail applying the BSPCloud library on matrix multiplication. The performance predictability and speedup are evaluated in the cloud platform. We give the results of experiments.

Cloud computing integrates vast computing and/or storage resources together, which provides services on demand via network. The cloud computing data center is usually composed of thousand of commercial computers, and these computers are connected by network. A cloud computing programming model is urgent to design to help user to use cloud resources without concerning the details of implementation. Nowadays, some cloud computing programming models have been proposed. This paper first analyzes the problems which cloud computing programming model need to solve, and then analyzes the characteristics of the cloud computing programming model. The advantages and disadvantages of each programming model are proposed. Finally, this is paper concludes the current issues and future trends.

It is difficult to deeply develop computing ability of supercomputer through programming. There are two approaches to resolve this problem. The first is that programmer masters advanced knowledge about architecture of the supercomputer. The second is to construct a friendly programming environment to ease the difficulties in parallel programming. Julia is a scripting language which supports language-level high performance computing. It is easier to implement parallel algorithm in Julia language. So, we develop a programming supporting environment for Julia to edit and debug source code, and prepare a set of portable function library for heterogeneous supercomputers. These programs are sent to supercomputers through a message system. As job description, these Julia codes are scheduled to proper nodes and executed based on the portable function library. In the computing case of bus-line statistics, the parallel Julia program on 4 nodes achieved a speed up of 3.15x.

Programming models for cloud computing has become a research focus recently. Cloud computing promises to provide on-demand and flexible IT services, which goes beyond traditional programming models and calls for new ones. Some progress has been made in cloud computing programming models for large-scale data processing, but little was done on models of predictable performance. With the advantages on predictable performance, easily programming and deadlock avoidance, the BSP model has been widely applied in parallel databases, search engines, and scientific computing. This paper targets to adapt the BSP model into cloud environment. The scheduling of computing tasks and the allocation of cloud resources will be integrated into the BSP model. A BSPCloud programming model with predictable performance is proposed.

클라우드 컴퓨팅은 가정과 비즈니스 컴퓨팅에 있어 빠르게 확산되고 있다. 이러한 변화에 따라 수년 내에 다양한 클라우드 서비스를 제공하는 클라우드 서비스 공급자의 수가 폭발적으로 증가할 것으로 예측된다. 하지만 클라우드는 하나의 독립체가 아니고 서로 단절되어 있는 많은 응용 프로그램(SaaS)과 기반시설(IaaS), 상이한 플렛폼 서비스(PaaS)들의 집합으로 볼 수 있다. 클라우드 중개방법(brokering mechanism)은 산재되어 있는 서비스들의 집합을 통합하고 편리하게 유용한 서비스를 다양한 사용자에게 제공할 수 있게 하기 때문에 서로 상이한 클라우드 시장을 하나의 통상적 서비스로 변환시키는데 있어 필수적이다. 따라서 CSBs(클라우드 서비스 중개자들)는 사용자와 CSP(클라우드 서비스 제공자)사이의 거리를 최소화하고 예약과 on-demand 서비스의 비용을 최소화하는 운영전략을 찾을 필요가 있다. 이를 위하여 본 연구에서는 클라우드 서비스의 비용과 품질을 최적화하기 위해 예약 접근법을 고려한 수리적 모형을 제안한다.

Cloud computing is fast becoming the wave of the future for both home and business computing. Because of this growing acceptance, we can expect an explosion of diverse cloud service providers in the coming years. However, the cloud is not a single entity, rather it is a set of many disconnected islands of application (SaaS), infrastructure (IaaS), and different platform (PaaS) services. Cloud brokering mechanisms are essential to transform the heterogeneous cloud market into a commodity-like service. Cloud service brokers (CSBs) are the new form of business entities to help to aggregate the scattered set of cloud services and make conveniently available to their diverse users. CSBs can reserve a certain percentage of their clients' (users') demand and satisfy the remaining portion in an on-demand basis. In doing so, they need to minimize cost of both reserved and on-demand instances as well as the distance of a link between the cloud service provider (CSP) and the user. The study proposes a reservation approach with a mixed integer model to optimizes the cloud service cost and quality.

컴퓨터 프로그래밍 실습을 위한 가상 데스크탑 서비스를 제공하기 위해서는 각 교과별로 사용자 그룹이 만들어져야 하며, 개발도구, 디스크 이미지, 사용자 계정 정보, 로그 데이터 등을 관리하기 위한 관리 프로그램이 필요하다. 본 논문에서는 대학에서 컴퓨터 프로그래밍 실습 교육에 활용할 수 있는 클라우드 컴퓨팅 기반의 가상 데스크탑 서비스 제공 방안과 효율적인 운영 관리 방안을 연구하였다. 구현된 가상 실습 환경 운영 관리 시스템을 이용하면 각 교과의 커리큘럼에 적합하게 커스토마이징 된 실습 환경을 사전에 미리 구축하여 교과별로 빠르게 프로비저닝 할 수 있다.

In order to provide the virtual desktop service for computer programing training, User group to use the service should be created for each subject. And the management program to manage development tools, disk images, user account information, log data is required. In this study, we implemented the web-based operation management system to manage the virtual desktop service for computer programming training. We can provide rapid provisioning for the virtual desktop service by using the implemented system.