Gemini Open Cloud Solutions Version 6 4/2016 Make Cloud Computing Easy for Business! Gemini Open Cloud Computing Inc 11F-3, No. 158, Sec 2, Gongdao 5 th Road, Eastern Distrct Hsinchu City, 30070 Taiwan Tel: +886-3659-0698 Fax: +886-3575-3186 http://www.geminiopencloud.com
Table of Contents Gemini Open Cloud Solutions... 1 Version 6 4/2016... 1 Introduction... 1 Video Surveillance Cloud Solution... 2 Hadoop Solution... 4 Gemini Engineering Cloud Solution... 8 BackOffice Cloud Solution... 11 DevOps Cloud Solution... 13 Gemini Open Cloud Solutions ii
Gemini Open Cloud Solutions Version 6 4/2016 Introduction Gemini Open Cloud (GOC) is an end-to-end cloud infrastructure platform that enables a service provider to convert legacy applications into cloud services without modifying the application binaries. Gemini Open Cloud (GOC) enables a CSP to manage the Life cycle of a Cloud Service including Cloud Service definition, Provisioning & Orchestration, Customizing a management portal, and provides monitoring and reporting on the cloud service resource usage. GOC comes with an all-in-one virtualization infrastructure, plus the GOC PaaS platform to orchestrate, manage, and monitor its hosted Cloud Services. Through the GOC solution foundry, cloud service providers can add a new cloud service to GOC with just a few clicks and customize a service portal for its administrators and end users. Once the cloud service is up & running, the GOC PaaS platform monitors the resource utilization and pre-defined events. GOC also generates historic resource usage reports on a per service or per user basis. The following sections details some of the example cloud services we have integrated on the GOC platform. Please note, however, GOC is not limited to such services, Gemini has the professional service and SI partners to help ISV to port any legacy application onto GOC and operate them as cloud services. Gemini Open Cloud Solutions 1
Video Surveillance Cloud Solution The widespread replacement of traditional CCTV by IP-based cameras has enabled the transfer and storage of video content to the cloud. The video can be monitored live or recorded in cloud storage. This service is called Video Management Services (VMS). Within the cloud, a set of cloud servers can process such recorded videos for summarization, pattern recognition, search and analysis, central console monitoring, event notification, and much more. This is called Intelligent Video Services (IVS). Gemini Open Cloud Computing has long worked on VMS and IVS when it was still part of CCMA in ITRI (Industrial Technology Research Institute). Leveraging our GOC PaaS platform, we can provide the following benefits to IP-based video management: Resource pooling Server, Storage, Networking resources can be virtualized and shared by multiple tenants, each owning their set of IP-based cameras. Anywhere device network The IP cameras can practically be located anywhere. Gemini Open Cloud can scan for these cameras and manage them. Videos will be transferred over WAN to be stored and analyzed. Elasticity compute resources Video analysis can be CPU intensive. Leverage the virtual machines; we can pull together computing resource when needed to do the analysis. Such resources can dynamically be allocated for tenants as needed. Elastic storage resources The storage of IP-based video in the cloud can eliminate the issue with limited storage capacity in the camera, and take advantage of cloud storage features such as automatic data replication, thin provisioning, de-duplication, and accessible anywhere. Measured Service The best part is tenants are only charged by the resources they used, rather than expensive capital equipment they have purchased based on maximum capacity Gemini Open Cloud Solutions 2
High Availability Our cloud infrastructure will provide server failover and data replication, so the IP video management services are accessible on-demand, around the clock for tenants. In short, the Gemini Open Cloud video surveillance solution makes it simple for carriers & service providers to provide advanced IP Video Management. Traditional IP Video Management software takes hours, if not days, to set up. With automatic camera and server discovery - and Gemini s orchestration layer, it will only take minutes to install and configure such services. The service providers can request additional computing and storage resources on demand. This eliminates the need for them to go through the capital budget process to acquire additional hardware. For the end users, they will only have to pay for what they use, although they can access the service with any device, anywhere in the world. Gemini Open Cloud Solutions 3
Hadoop Solution Enterprises are increasingly recognizing the importance of Hadoop to their data strategy success. A recent TDWI (The Data Warehouse Institute) survey found that only 34 percent answered that Hadoop was not a pressing issue. Over 40% of the respondents expects to have Hadoop/HDFS in production in the next 12 months, and also over 40% of the respondents feel that Hadoop will benefit their company s big data collection and advanced analytics. It's evident there's a market out there for Hadoop. The question is how you benefit from combining Hadoop and Cloud by running your Big Data acquisition and Map Reduce Data Analysis job on a platform like Openstack. Let s first examine what s the benefit of virtualizing the Hadoop components. Gemini Open Cloud Solutions 4
Why virtualize Hadoop nodes? o Increased performance: Virtualized Hadoop nodes have been proven to perform better than their bare-metal cousins4 o Hardware utilization: Bare-metal Hadoop deployments average 10-20% CPU utilization, a major waste of hardware resources and datacenter space. Virtualizing Hadoop allows for better hardware utilization and flexibility o Elastic Map Reduce and scaling: Dynamic addition and removal of Hadoop nodes based on load allow you to scale based upon your current needs, not what you expect. Enable supply and demand to be in true synergy o Allow DevOps & IT Ops to live in harmony: Big Data scientists demand performance, reliability, and a flexible scale model. IT Ops relies on virtualization to tame server sprawl, increase utilization, encapsulate workloads, manage capacity growth, and alleviate disruptive outages caused by hardware downtime. By virtualizing Hadoop, Data Scientists and IT Ops mutually achieve all objectives while preserving autonomy and independence for their respective responsibilities o Sandboxing of jobs Make Hadoop and Enterprise Apps play nice: Buggy Map Reduce jobs can quickly saturate hardware resources, creating havoc for remaining jobs in the queue. Virtualizing Hadoop clusters encapsulates and sandboxes Map Reduce jobs from other important sorting runs and general purpose workloads o Batch Scheduling & Stacked workloads: Allow all workloads and applications to co-exist, e.g. Hadoop, Virtual Desktops and Servers. Schedule Map Reduce job runs during off-peak hours to take advantage of idle night time and weekend hours that would otherwise go to waste or utilize VMware s resource pooling features to run concurrently o New Hadoop economics: Bare metal implementations are expensive and can spiral out of control. Downtime and underutilized CPU consequences of physical server s workloads can jeopardize project viability. Virtualizing Hadoop reduces complexity and ensures success for sophisticated projects with a scale-out grow as you go model a perfect fit for Big Data projects Gemini Open Cloud Solutions 5
o Service-defined tiering: Maintain and manage SLAs with resource prioritization and reservations The GOC Hadoop Solution provides the following services: Create/Delete Hadoop Clusters User can create/delete a Hadoop cluster with arbitrary number of nodes (VM s) and install your favorite Hadoop distro in this cluster (Apache, Clouderra, or HortonWorks). Scale up on Hadoop Cluster Additional member nodes (VM s) can be added to an existing cluster to run new jobs. Submit a Hadoop Job A user can upload a Map Reduce job + input data to an object storage container to launch a job in a previously created Hadoop cluster. Similarly, the job output can also be placed in a persistent object storage container, so when the job is completed Gemini Open Cloud Solutions 6
and the VM s are deleted, the user can still access the storage container to look at the console logs and output. Cancel a Hadoop Job Both a running job and a scheduled job can be cancelled from the Hadoop job service portal. Re-launch Hadoop Job A user may want to run his Map Reduce job with different input parameters. In such case, the user only need to upload his new configuration file to the input storage container and then re-launch the job. Typically, a Hadoop Map Reduce job perform data mining from a data warehouse stored in HDFS. Because HDFS allows the addition of data nodes dynamically, the GOC Hadoop service allows infinite growth of the data warehouse (by spawning additional VM s) as more and more data are extracted from the data source into HDFS. Realizing the data source can come in different repository and storage medias, Gemini Professional service will also work with customer on different ETL scenarios to extract data from different data source, by leveraging the rich set of tools from the Hadoop ecosystem. Gemini Open Cloud Solutions 7
Gemini Engineering Cloud Solution Traditional Engineering applications like Matlab, Ansys, SPSS, and Pro-E are purchased by individual organizations within an institution to run on dedicated workstations. This approach limits the hardware capacity to that of the physical workstation and there is very little sharing between organizations within an institution. The Engineering Cloud Solution powered by Gemini Open Cloud is the perfect solution to transform these legacy applications into Services hosted on our elastic cloud infrastructure, without changing the application itself. Through Gemini Open Cloud s orchestration layer and management portal, the end users only have to click a few buttons and easily create an instance of engineering service like Matlab or Ansys, without the hassle of acquiring hardware and application licenses. Furthermore, a user can request a cluster of virtual machines to run such engineering applications in a scale out architecture (cluster mode). Most applications already support running their engineering applications in a cluster configuration. When the research is completed, the VM resources can be relinquished for other users. GOC s Engineering cloud solution not only promotes sharing of physical machine resources, it also empowers the sharing of different toolkit licenses to the entire constituent of the institution. Traditionally, these toolkits are very expensive and it s unlikely an individual organization can afford all toolkits for its departmental users. As a result, individual users will have to spend much more time to implement the algorithm and there are a lot of duplicate efforts across organizations. Gemini Open Cloud Solutions 8
Because the individual users access these virtual machines through remote desktop interface, the users do not have to learn any new UI s to use such cloud services. The users can access these applications anywhere, anytime, on any devices. For heavy graphic rendering applications, Gemini also supports virtualized GPU s to attach to these virtual machines to optimize the frame buffer data transfer between the VM and the desktop. Beside managing virtual resources and licenses, GOC also provides simple, yet powerful portal for service providers as well as individual users to monitor the usage and status of their virtual resources. Gemini portal will generate detail reports based on scope and authority of the currently logged on user. The raw data of these report can be fed into billing system for the institution s IT department to charge the individual organizations based on their usage, rather than acquiring the licenses on a perpetual basis. The Gemini Engineering Cloud Solution has been in production usage by large institutions and has been proven to be an effective solution based on its ROI. Further data is available upon request. Gemini Open Cloud Solutions 9
Gemini Open Cloud Solutions 10
BackOffice Cloud Solution Back office applications are those that run on the backbone infrastructure of an office environment. These include Content Management, Document Management, VDI, mail, meeting makers, web based conference, Enterprise instant messaging, CRM, ERP, and even HR, & payroll software systems. Because of security and data sensitivity, these applications used to run on dedicated servers and access is strictly controlled. Therefore, they are financially a heavy burden for Small and Medium business. The multi-tenancy and isolation (via virtualization) concept of cloud is a perfect candidate to allow multiple SMB s to share the same cloud infrastructure and yet not compromise on data security. Gemini Open Cloud currently offers a Content & Document Management service that is based on best of breed open source software packages. We use Alfresco to power collaboration between multiple content contributors within the Enterprise. The concept of mini-sites and different role hierarchy allows contents to be partitioned and firewalled between organizations of an Enterprise. The workflow engine enables reviews and only authorized contents to be published. Approved web content can be published internally or externally via open source web servers such as Wordpress or Drupal. Gemini Open Cloud also integrates with other ISV vendors to provide paperless meeting coordination, online meeting collaboration, and enterprise cloud storage application. For details, please contact our marketing department for a demo. Gemini Open Cloud Solutions 11
Gemini Open Cloud Solutions 12
DevOps Cloud Solution Continuous Integration is becoming the norm for software development in recent years. As a firm believer of CI process, Gemini has integrated a set of tools on our Gemini Open Cloud platform to aid software developers from Enterprise Java and Web application to mobile apps. What s CI Process? Continuous integration (CI) is the practice, in software engineering, of merging all developer working copies with a shared mainline (Release Integration Branch - RIB) several times a day. The sub-tasks involved in CI include (but not limited to): 1. Maintaining a code repository (Github, CVS, SVN, Perforce ) 2. A code review process upon all commits 3. Make it easy to do Get Latest and provide tools for merging and resolving conflicts 4. Automating the build process 5. Make the build self-testing 6. Require developers to commit daily and do nightly automated builds/testing 7. Doing integration test on a clone of the production environment. Gemini Open Cloud Solutions 13
Gemini Open Cloud integrates the most popular open source development tools such as Eclipse, Github, Review Board, Jenkins, Robot, Selenium, Sikuli and test case management tools such as Spira, as well as problem tracking tools such as bugzilla onto a common cloud platform, so developers can do end to end development, and collaborate with testers to resolve problem quickly. Dashboards are available for all developers to share the build and testing status. Using the Gemini Open Cloud DevOps solution, software developers can practice iterative test-driven development. The development community will be much more collaborative, yet precise in the problem tracking and resolution process. The organization can truly accomplish continuous delivery of small increments until the final product that meets all customer requirements. Gemini Open Cloud Solutions 14