- Client and server
- Peer to peer
- Hybrid of client server
Host-based Architectures - everything is in one machine(mainframe)
Problems with Host-based Architecture
- Host becoming a bottleneck : Everything is on a server, server has to do all processing
- Upgrade typically expensive and "lumpy"
Client-based Architectures
- Client need a high performance
- Server keep data but no processing
- When client want information of someone, Server can send just whole data and Client need to find by themself
Client-Server Architectures (Used by most network today)
- Client no need very high performance (don't have to spend a lot of money)
- When Client want information of someone, Server can send just that information to Client
Advantage of Client-Server Architectures
- Advantages
- More efficient because of distributed processing
- Allow hardware and software from different vendors to be used together
- Disadvantage
- Difficulty in getting software from different vendors to work together smoothy
- May require Middleware, a third category software
Fat vs. Thin Clients
- Fat clients do a lot of work in client based Architecture
- Thin client is an easy machine(low cost machine) which can control user not to use unlicensed program
Criteria for choosing Architecture
- Infrastructure Cost : Cost of hardware
- Development Cost : Develop on mainframe, cheaper than develop on Client, Server
- Scalability : Can add new server is cheaper cost than mainframe
3-Tier Architecture
- Client (microcomputer)
- Application server (microcomputer)
- Database server (microcomputer, minicomputer or mainframe)
- Application server and Database server are divided to do part of logic but both of them don't have to do much work
N-tier Architecture
- Data have to travel along network from client side to server side
- Database server can divide in more than 1 tier (having another tier), if Database server reply something, it needed to pass to every server
Multi-tier Architectures
- Advantages
- Better load balancing
- More stable(if server work hard, just add more server)
- Disadvantages
- Heavily loaded network
- Difficult to program and test due to increased complexity
DNS - Domain Name Services - Use to translate URL (DNS) to IP address
How DNS works
1. Assume that Susie asks for a web page on Indiana University's server by access in www.susie.indiana.edu from University of Torronto
2. Client will ask DNS server that know Susie or not
3. If no, Ask Root that know Susie or not
4. If no but know Indiana, Send request to Indiana
5. Then send result back to Susie at Torronto
ไม่มีความคิดเห็น:
แสดงความคิดเห็น