Solution Brief:
Development Kits Based on Intel® Architecture for Intel® Core™ Processors
Intel® Celeron® Processor
"With the rise of edge computing, more and more users hope to develop and verify applications such as edge AI at the edge through lightweight devices such as single-board servers. In this process, a healthy developer ecosystem The establishment of edge computing is crucial. Developers are the main force in promoting edge computing technology innovation and market applications, and the energy of developers is the cornerstone of edge computing development. ZimaBlade, ZimaBoard and ZimaCube of IceWhale Technology based on Intel® processors are developers It provides unimaginable scalability, performance and reliable security, and has become an important carrier for developers and geek users to unleash their imagination and unleash their innovative potential."
—Lauren Pan
Founder of Ice Whale Technology
Overview
For "geek" users and enterprises who want to flexibly expand hardware and software to meet the needs of different application scenarios, obtaining a simple single-board computer (SBC) is very attractive. This type of product can act as a micro-server and be used in scenarios such as network-attached storage (NAS), virtual private network (VPN), software routers, media servers, etc., providing lightweight load support capabilities and high scalability. and flexibility. But at the same time, users’ innovation needs in areas such as artificial intelligence (AI) mean that SBC needs to keep pace with the times in terms of computing power and software ecology to support a wider range of applications.
ZimaBlade, ZimaBoard and ZimaCube are low-cost SBC products designed by IceWhale Technology for manufacturers and geeks. It combines the high scalability of x86 SBC with the low power consumption and small size of a micro server. These three products are equipped with Intel® Celeron® processors and Intel® Core™ processors respectively, which can give full play to the advantages of processors in low power consumption, high performance, high IO expansion and high stability. At the same time, these two processors can also enable more application scenarios through the Intel Iris® Xe graphics card, combined with the powerful functions provided by Intel® oneAPI video processing library (Intel® oneVPL) and the OpenVINO™ tool suite.
challenge
SBC is a complete computer built on a single circuit board. It usually includes core components such as processor, memory, storage, input/output (I/O) interface, etc., and provides basic computing, storage, network and other capabilities. SBC generally has strong scalability and supports flexible customization. Currently, SBC is usually used as a demonstration or development system and is widely used in product prototype design, personal computing system construction, industrial edge computing and other scenarios.
The growing developer community is a huge driving force for the growth of the SBC market, as shown in the "China Developer Portrait Insights Research Report" released by InfoQ , China's developer group will reach 20.1637 million in 2022, and the average annual group size growth rate since 2016 has exceeded 10%. Among them, front-end and back-end developers account for the largest proportion, reaching 8.6 million. With the expansion of the scale of developers and the continuous improvement of SBC chips in terms of computing power and energy consumption, the SBC market has achieved rapid growth. Another report shows that the global and Chinese SBC market sizes will reach 5.693 billion yuan (RMB) and 1.683 billion yuan respectively in 2022. The global SBC market size is expected to reach 12.0 billion yuan by 2028, with a compound annual growth rate (CAGR) of 13.02%.
From the perspective of market demand, the widespread application of SBC still needs to overcome the following challenges:
- How to provide sufficient computing power to support the efficient operation of conventional and new workloads
SBCs are usually small in size and have insufficient heat dissipation conditions, so they have strict requirements for chip power consumption. Main components such as CPU need to have a high energy efficiency ratio to provide sufficient computing power for workloads such as NAS and soft routing. At the same time, in the face of the increasing demand for supporting new workloads such as AI inference, more and more users hope that SBC can provide certain AI acceleration capabilities.
-
How to achieve flexible expansion to meet the needs of more scenarios
Rich scalability is the main feature of SBC, but it has also become a major challenge faced by SBC. Thanks to the complete ecosystem of x86 architecture, SBCs based on x86 architecture usually have more advantages in terms of scalability and compatibility, and can expand more peripherals and integrate more software applications. -
How to provide a rich software ecosystem and support the rapid construction of target systems
To use SBC to build a complete system that can be put into use, in addition to hardware preparation, software applications also need to be prepared. Very important. Finding matching applications and images is usually tedious and requires repeated debugging and verification.
Solution: IceWhale Technology Zima Series Products Based on Intel® Architecture
Icewhale Technology has launched ZimaBlade, ZimaBoard, ZimaCube and other products based on x86 architecture. Among them, ZimaBlade is equipped with Intel® Celeron® processor, which is a low-power, flexible, modular, palm-sized and multi-core server that redefines the standard of lightweight servers. This product can be used to build long-tail application scenarios such as micro servers, light NAS, smart homes, gateways, etc. It supports users to build exclusive and secure data asset interaction hubs to easily access, share and synchronize files, contacts, passwords, etc., improving data security. In addition, ZimaBlade can also be used as a streaming media server to stream videos, pictures, music and other files.
Intel® Celeron® processors have excellent performance per watt, rich graphics effects, and integrated I/O capabilities to provide outstanding processing power and support for a variety of image sensors. This processor delivers the low-power computing performance needed for small designs and applications, while maintaining smooth application responsiveness while browsing and multitasking. By using Intel® Celeron® processors, ZimaBlade can connect to x86's rich application ecosystem and run complex applications with long-term stability through the processor's excellent performance.
ZimaBoard is also powered by the Intel® Celeron® processor, which delivers superior video conferencing capabilities, faster wireless connectivity, improved overall app and graphics performance, and long battery life. ZimaBoard features modular design, low power consumption, low cost, and small size. It can use modular hardware and software options to facilitate users to flexibly build application systems such as NAS, VPN, soft routing, media servers, smart homes, and streaming media. . Under the creative stimulation of the developer group, a variety of new applications and product forms can also be formed.
ZimaBoard supports accelerated AI inference through the OpenVINO™ tool suite, which is used to quickly develop applications and solutions to solve a variety of tasks (including human vision simulation, automatic speech recognition, natural language processing, recommendation systems, etc.) Comprehensive tool suite. Based on the latest generation of artificial neural networks, including convolutional neural networks (CNN), recurrent networks, and attention-based networks, this tool suite scales computer vision and non-vision workloads across Intel® hardware, resulting in significant performance improvements. The OpenVINO™ tool suite accelerates applications with high-performance, artificial intelligence and deep learning inference deployed from edge to cloud. It can accelerate AI inference on components such as CPU and integrated graphics card, and can replace independent AI inference accelerators in some loads [Note: The data is quoted from the test results of Icewhale Technology as of August 2023. Test configuration: Zimaboard 832, Intel® Celeron® Processor N3450 @ 1.1 GHz, Intel® HD Graphics 500, 8 GB RAM, 32 GB HDD, CasaOS 3.38.5 64-bit, OpenVINO 2023.0, using Intel Model Zoo pre-trained model . Intel does not control or audit third-party data. You are asked to review this content, consult other sources, and confirm that the data mentioned is accurate].
Icewhale Technology test data shows that when performing biometric recognition on 1080P video, the CPU (Intel® Celeron® processor) + OpenVINO™ tool suite can achieve a real-time AI inference processing speed of 13 FPS, and the integrated graphics card (Intel Iris) ® Xe graphics card) + OpenVINO™ tool suite can achieve a real-time AI inference processing speed of 21 FPS.
[Note: The data is quoted from Icewhale Technology’s test results as of August 2023. Test configuration: Zimaboard 832, Intel® Celeron® Processor N3450 @ 1.1 GHz, Intel® HD Graphics 500, 8 GB RAM, 32 GB HDD, CasaOS 3.38.5 64-bit, using Intel Model Zoo pre-trained model. Intel does not control or audit third-party data. Please review this content, consult other sources, and confirm that the data mentioned is accurate. ]
The product also supports calling the Video Acceleration API (VAAPI) for hardware acceleration, realizing hardware acceleration support for integrated graphics card encoding, decoding and transcoding, making ZimaBoard faster when processing video-related requests. VAAPI is an open source library and API specification that supports calling the hardware acceleration capabilities of graphics cards for video processing. It consists of a main library and driver-specific acceleration backend libraries. These backend libraries target every supported hardware vendor.
Icewhale Technology conducted a video (H264, MP4 format, 30fps) transcoding and decoding performance test. The test data is shown in Figure 4. After the transcoding and decoding load is offloaded to the integrated graphics card and supported by VAAPI, the performance of the integrated graphics card can be fully utilized. Compared with the CPU (software decoding), the transcoding and decoding performance can be improved to About 3.39 times the original one. [Note: The data is quoted from Icewhale Technology’s test results as of August 2023. Test configuration: Zimaboard 832, Intel® Celeron® Processor N3450 @ 1.1 GHz, Intel® HD Graphics 500, 8 GB RAM, 32 GB HDD, CasaOS 3.38.5 64-bit, FFmpeg 4.3.6, LibVA 2.15.0, VAAPI 1.15.0. Intel does not control or audit third-party data. Please review this content, consult other sources, and confirm that the data mentioned is accurate. ]
[Note: The data is quoted from IceWhale Technology’s test results as of August 2023. Test configuration: Zimaboard 832, Intel® Celeron® Processor N3450 @ 1.1 GHz, Intel® HD Graphics 500, 8 GB RAM, 32 GB HDD, CasaOS 3.38.5 64-bit, FFmpeg 4.3.6, LibVA 2.15.0, VAAPI 1.15.0. Intel does not control or audit third-party data. Please review this content, consult other sources, and confirm that the data mentioned is accurate.
]
ZimaBoard can flexibly meet the needs of the following scenarios:
- Personal Cloud: ZimaBoard allows users to easily expand to 36 TB of disk space (2 x SATA) [Note: Data is quoted from IceWhale Technology’s internal testing results. Intel does not control or audit third-party data. Please review this content, consult other sources, and confirm that the data mentioned is accurate. ] to meet the need for more data storage. When connected via a Gigabit network, users can enjoy a fast read and write experience just like a local hard drive.
- Software router: With the excellent compatibility provided by the x86 chipset, users can turn ZimaBoard into a hardware router, create secure VPN connections, run lightweight services such as FTP servers and BitTorrent clients, and perform traffic shaping and service level agreements (QoS). ) and even set up private access to your work computer.
- Media Server: Users can manage all their media, stream favorite movies, TV show collections and enjoy high-quality tracks anywhere on all their devices using the media server installed on ZimaBoard.
- File Sharing and Collaboration: Users can use ZimaBoard to seamlessly enhance teamwork on documents and spreadsheets in a protected environment. Users can set up large-capacity shared disks at home or in the office to edit videos, photos and documents with others in real time on multiple systems.
- Personal Server: ZimaBoard offers excellent development possibilities as a personal server and can be used as a cost-effective x86 microserver. For example, it is ideal for deploying Git servers, web servers (Apache/Lighttpd/Nginx), etc.
- Industry/embedded projects: Users can use ZimaBoard for data collection, data buffering and data pre-processing to provide accurate and effective on-site IoT data to cloud or local servers, significantly reducing dependence on cloud and network infrastructure.
For users with higher performance requirements, IceWhale Technology provides ZimaCube equipped with Intel® Core™ processors, which has more powerful computing power, higher-end hardware configurations and more complex designs, and can meet the needs of media servers, productivity Requirements for building tools, data centers, private AI, rendering servers, etc.
ZimaCube supports the 12th generation Intel® Core™ processors, which redefine multi-core architecture for edge and end devices with a new performance hybrid architecture. The Performance Core (or "P Core") dramatically improves single-threaded performance and responsiveness, while the Energy Efficiency Core (or "E Core") delivers scalable multi-threaded performance and efficient background task offloading for modern multitasking. The processor includes Intel Iris® Xe graphics powered by Intel®
ZimaCube can also use Intel® oneVPL to accelerate the processing of video data. Intel® oneVPL features easy-to-use video processing software development tools and complete pipeline demonstrations to dramatically accelerate video playback, encoding, processing, and media format conversion. Intel® oneVPL enables prototyping, optimizing, and productizing media pipelines using comprehensive and convenient APIs to accelerate time to market. Intel® oneVPL can implement video encoding and decoding hardware acceleration on Intel hardware platforms, and supports new encoding optimization technologies such as Low Latency Encoding. The solution also provides support for AV1 on the latest Intel® hardware, which can bring users higher compression rates, better performance and lower bandwidth consumption.
ZimaCube can play an important role in the following scenarios:
- Private Cloud
ZimaCube can support users to build ten-core, 6-disk [Note: The data is quoted from the internal test results of Ice Whale Technology. Intel does not control or audit third-party data. Please review this content, consult other sources, and confirm that the data mentioned is accurate.
], a personal private cloud that supports multi-channel solid-state drive (SSD) expansion. It has a modular design, multiple PCIe 3.0 channels and an open private cloud application system. -
Local AI Assistant
ZimaCube is compatible with cloud disks, local storage and collaborative applications, and supports synchronization of all data to the user's own personal private cloud. ZimaCube also supports organizing all information through the local AI assistant, helping users find the personal information they need through Chat conversations, and efficiently assisting in the secondary creation of content. -
Media editing collaboration and sharing
ZimaCube can be used to build a media editing collaboration and sharing platform, supporting multi-person remote and studio material sharing and collaborative editing. Compatible with 1 GB/s data transfer and migration experience between Mac and SSD. -
Personal AI Cockpit
ZimaCube can become a personal cockpit that supports hundreds of terabytes of storage space. It can access personal distributed data through local large models and become a personal local data steward. . It has a modular design, GPU computing power-compatible design and an open private cloud application system, making it an ideal choice for personal AI cockpits.
The Zima range also offers easy access to CasaOS. CasaOS is an open source smart home operating system that aims to provide users with a simple, flexible, and secure home cloud platform that can support users to freely install various APPs to realize a variety of smart home scenarios. CasaOS is a community-driven open source project, and everyone involved is also a user of CasaOS, enabling every developer and user to become a promoter of its development.
Based on its software and hardware products, Icewhale Technology is committed to connecting industrial scenarios/end consumers/developers and building a prosperous ecosystem. It not only uses software as a tentacle to meet the personalized needs of life/production, but also quickly and timely Long-term operation to solve the edge computing needs of small and medium-sized enterprises, support small and medium-sized enterprises to solve pain points and explore innovative directions under the wave of AI, and ultimately create a spiral growth model with high stickiness, ecological penetration, and technological prosperity.
As a computing platform for localized large models, the Zima series of products can form a variety of new applications and product forms under the creative stimulation of the developer community. For example, the ZimaBoard product is currently used by the first group of geeks to explore a variety of application scenarios, such as firewalls, soft routing, etc., and has formed a large number of successful cases, providing a reference for other users to explore.
income
In order to promote the construction of a universal edge computing platform ecosystem, for "geek" users, the Zima series products provide a low-cost, highly flexible basic platform. Through the combination of modular hardware, users can freely implement various innovative applications and product form, thereby stimulating their creativity. On the other hand, it provides a flexible development and verification platform for enterprise users, supporting them to flexibly customize according to market needs, laying a solid foundation for the industrial development of products and promoting comprehensive ecological productivity.
The Zima series of products can bring the following values to users:
- Compatible with a large number of mainstream operating systems and application ecosystems, and low-cost access to applications: Thanks to the x86 architecture, ZimaBlade, ZimaBoard and ZimaCube are compatible with mainstream operating systems, a large number of peripherals, and rich applications, making it easy to implement functional expansion quickly and flexibly. Application needs of different scenarios.
- Efficiently run target applications while saving energy and controlling size: ZimaBlade, ZimaBoard and ZimaCube have the characteristics of low power consumption and high performance, and can run loads such as data storage, data forwarding, multimedia processing and even AI inference with small-sized devices. It helps to save investment and reduce investment in cooling, power supply and other aspects.
- High stability and high availability enable continuous product operation:
The Intel® Celeron® processors and Intel® Core™ processors used by ZimaBoard and ZimaCube are both oriented to the Internet of Things and the edge Computing scenario optimization, with long-term proven stability and availability, ensures the product's continuous operation capabilities in various scenarios. - Stimulate vitality through an active community ecology:
Icewhale Technology has built an active user ecology, with rich community cases of Casa and Zima. Users can use the power of the ecology to enrich products Application scenarios inspire product value.
In secure access scenarios, Icewhale Technology’s Zima products also have huge application potential. Many enterprise branches have problems such as weak network security capabilities, lack of full-time operation and maintenance personnel, difficult network access maintenance, and difficulty in network access between branches and headquarters. Some enterprises access internal systems through VPN, which poses certain security risks and the dedicated line costs are relatively high. high. The solution based on ZimaBoard of Icewhale Technology is mainly used for POP (Point-of-Presence) nodes in actual business scenarios such as enterprise multi-branch networking, remote office, enterprise security reinforcement, and Internet of Things security protection, using SDN and SD-WAN. The feature of dynamically programmable network configuration improves network performance and management efficiency. At the same time, ZimaBoard has high availability and can improve the sustainable operation capabilities of large-scale network equipment. This solution can flexibly customize network functions on the controller or upper-layer applications, realize centralized deployment and management of security policies, and better meet the needs of enterprise networks.
Outlook
In addition to having a broad market among developers and geek-type users, Icewhale technology products also have extensive application potential in the enterprise-level market of edge computing. The company provides a platform with low power consumption, high computing power, high scalability, and high stability that runs on the edge and can handle large-scale data and complex tasks. In addition, these products can be flexibly configured according to customer needs, allowing mid- to long-tail customers to choose the functions and services they need based on their own needs, while keeping deployment and management costs low.
Icewhale Technology will further promote the industrial development of Zima series products, including smart home, industrial automation, entrepreneurial industry, intelligent transportation and other fields, making the Zima series products an important carrier for edge AI applications, helping customers meet real-time needs, Protect data privacy and security, reduce network bandwidth requirements, provide offline capabilities and autonomy, and enable large-scale deployment.
In the future, Ice Whale Technology will further expand the base of players and developers by launching more cost-effective micro servers. At the same time, CasaOS will also provide better software functions around home data management to better meet the home and remote collaboration needs of users ranging from content creators, designers to audio and video practitioners. At the same time, Icewhale Technology will also strengthen cooperation with Intel in the field of SBC, including the use of new generation Intel® processors, OpenVINO™ tool suite, Intel® oneVPL and other software technologies, to provide a richer range of product and solution options to help users achieve success. Educate makers to connect with industrialized applications to unleash more users’ innovation potential.
About Intel
Intel (NASDAQ: INTC) is an industry leader creating world-changing technologies that advance global progress and enrich lives. Inspired by Moore's Law, we continue to advance semiconductor design and manufacturing to help our customers solve their most significant challenges. By integrating intelligence into the cloud, network, edge and various computing devices, we unlock the potential of data to help make business and society better. For more information about Intel innovation, please visit the Intel China Newsroom at newsroom.intel.cn and the official website at intel.cn.
About Ice Whale Technology
Icewhale Technology was established in August 2020. It takes European and American consumer-level private clouds as the entry point, starting from the cloud edge tools of geek groups and creators, and is committed to serving 400 million global families and creating a super node for data collaboration management.