Hybrid client
From Wikipedia, the free encyclopedia
Hybrid client is a client/server network computing architecture where clients (end users computers), process locally but do not persist data. Responsibility for the computing tasks is a combination (or a hybrid) of both client and server. The client CPU and RAM are utilized and responsible for the processing, while relying on the server to take care for the storage. The typical client used in such networks is a diskless computer capable of network booting. Server side usually features central storage and network boot environment which streams operating system to the client.
The main motivation behind this concept is to minimize IT administration caused by local disks management, while still employing feature full clients. Or in other words, enjoy the benefits of both "thin client" and "fat client".
Contents |
[edit] Terminology
The term “hybrid” refers to the hybrid client/server responsibility or to the mixed characteristics of the fat and the thin client architectures. There are other terms that are often used to describe the similar architectures or techniques used to achieve such a network:
- "Network Computer". Network Computer was the first paradigm used to describe these type of computers. This initiative started in the mid 90's, but was not considered a commercial success. The production of network computers (compliant to the NC standards) stopped around the year 2000.
- "Diskless Workstation". While a thin client is also a diskless workstation, this term is usually used to describe local processing stations (identical to hybrid clients).
- "Network Virtual Disk". Refers to the central storage which acts as a virtual disk for the clients.
- "Thick client". This term is ambiguous. It can be either a PC with hard disk or without, and often referred to application architecture, and not network architecture.
- "Half-Thick client", "Chubby client", "Fit client". uncommon names for the same hybrid client.
- "OS streaming". The process of software delivery to the client.
[edit] Comparison with fat client
Both clients process locally, can handle peripherals and usually features high multimedia support. While fat client is an ideal choice for a single computer environments, it is hard to maintain a multiple fat clients network. Software install and management are done on each client separately and data is distributed. Enterprises often use several techniques to reduce administration cost, such as ghosting, whereas the software is cloned from one client to another. Even with cloning, each client is physically accessed and manually installed, causing high administration fees. In hybrid client networks the software is typically saved and managed in one central place such as a NAS and thus handled only once per change.
The reduced number of hard disks also reduces the TCO, saves power and lessen the chance for technical problems associated with hard disks. On theft there is no risk for data loss or industrial espionage because no information is persisted on the client. Diskless stations typically have no moving parts (a disk and a fan), which makes them reliable and also noiseless. This is ideal for certain purposes such as mission critical computers or media centers which should not interfere with audio.
Fat client on the other hand, has its own advantages:
- Bandwidth consuming. It does not consume any network bandwidth and does not rely on the presence of a working network connection.
- I/O performance. Disk operations on local hard disk are faster and have better latency.
Fat clients networks are also sometimes thought to be more stable because they offer more redundancy over hybrid networks. When a hard disk fails on a fat client, only one machine is not functioning, while disk fault on the central storage could lead to entire network breakdown. This is not accurate because central storage could be redundant itself (i.e. with multiple disks on a RAID).
[edit] Comparison with thin client
Both architectures employ diskless clients which have advantages over fat client (see above), but differ on the location of processing. This leads to few advantages for the hybrid client:
- Distributed load The processing load of the hybrid clients is distributed. Each user gets its own processing isolated environment, barely affecting other users in the network. Thin clients rely on the central server for the processing and thus require performant server which "works" for all users. When the central server is busy and slow, applications of all users will perform badly.
- Better multimedia performance. hybrid clients have advantages in multimedia-rich applications that would be bandwidth intensive if fully served. For example, hybrid clients are well suited for video gaming.
- More flexibility. On some operating systems (such as Microsoft Windows) software products are designed for personal computers that have their own local resources. Trying to run this software in a thin client environment can be difficult.
- Peripherals support It is considered much easier to integrate peripherals into hybrid clients.
Thin client advantages include cheaper hardware on the clients, and often better network bandwidth consumption.
