1. Classify the two alternatives in terms of what type of application architecture they use.
The first suggested solution is a client-based architecture since it’s the client that tells the server what data to select to analyze, making the client perform most of the work. The later architecture is a two-tier thin client architecture since it requires an application layer, in this case, a web browser and all processing is done by the server. 2. Outline the pros and cons of the two alternatives and make a recommendation to Fred about which is better.
Solution 1: Client-based pros-
1. Much more inexpensive to implement as micro-computers are considerably cheaper versus mainframe
2. Hardware …show more content…
of different manufacturers/platforms can function together on the same network
Solution 1: Client-based cons-
1. Data traffic must be sent and received between client and host
2. Due to the potential amount of information that must be sent between client and host, this can put a significant performance hit against the network, affecting all machines
Solution 2: Thin-client pros-
1. Better performance because processing is distributed to multiple machines (multiple web servers for example where load balancing is used to optimize network/server resources)
2.
Using a web browser as the client for this architecture means no deployment is necessary. Presentation is changed on the server and is immediately reflected on all clients accessing the server. Hardware and software from different vendors can be (is possible) used together
Solution 2: Client-server cons-
1. Getting software from different vendors to work together smoothly can difficult to setup. Not all platforms implementations of web browser are universal; something that renders in Internet Explorer may look different or not render at all if client is using FireFox.
2. Sometimes third party software may be required to translate between platforms via either Middleware or additional costs of development in adjusting client presentation to address the varying browsers and their implementations.
My recommendation is the thin-client solution. The server costs are minimal, allowing new servers to be added without affecting the existing infrastructure. Should a server go down and need replaced, a server can be removed, also without a significant impact on resources (assuming adequate servers existed to accommodate the amount of traffic experienced); using load balancing on the servers can help evenly distribute resource consumption. Finally, another advantage is using web browsers as clients. This reduces costs of having to deploy the application (addressing such issues as versioning), changes are reflected immediately when made on the
server.
Chapter 3: Mini-Case IV: Speedy Package
3. Assuming that each label is 1000 bytes long, how long does it take to transmit one label over the cell network, assuming that the cell phone network operates at 14kbps (14,000 bits per second and that there are 8 bits in a byte)?
Since we’re using an 8bit network, meaning there are 8 bits per byte giving us 8bits per byte * 1000bytes = 8000 bits. 14kbps can be re-written to 14000 bits per second so we can rewrite our equation as 8000 bit / 14000 bits meaning it takes .57142857 milliseconds (.571ms) to transmit a label @ 14kbps
4. If Speedy were to upgrade to the new, faster digital phone network that transmits data at 114Kbps (114,000 bits per second), how long would it take to transmit a label?
Since we’re still using an 8bit network which already determined is the same as saying 8000bits per label, change our divisor from 14000 to 114000.
8000/114000 = .070176436 milliseconds (.07ms) to transmit @ 114kbps