A deployment bag is part of the recovery system design to ensure successful parachute deployment. The use of a bag does not ensure successful recovery if the rest of the system is not designed and packed correctly. It is only one part of a system. A brief summary of what a bag should help do:
1. Orderly deployement. Get events to happen in the right order.
2. Get the parachute into clean air, away from turbulance that might cause a failure to open.
3. Get the parachute away from stuff (airframe, fins, nose cone, etc.) that might tangle it.
Using a bag is only one part of a successful recovery plan. There are other factors to consider, such as:
1. Protecting the recovery hardware (tubular nylon, parachute, etc.) from the BP charge.
2. Length of the riser ("shock cord").
3. Where the nose cone attaches to the system, if it is attached at all.
A deployment bag can be used with single deployment (pop the main at apogee) or dual deployment.
If a bag is used, how do you get the parachute out of the bag? Some bags are supposed to "fall open" and release the parachute. I've used bags that were supposed to do that. Sometimes they did. Sometimes they didn't. Everyone else (aerospace, skydivers, BASE jumpers, etc.) uses a pilot chute to pull the bag off the chute. The pilot chute has to produce enough drag to overcome whatever is holding the chute in the bag.
A rocket using dual deployment is usually falling fairly slowly, such as 50 - 75 feet per second. A skydiving falls at about 175 feet per second (120 mph). The slower an object falls and the less it weighs the larger the pilot to produce enough drag to separate chute and bag. I usually use a 36" pilot chute. If your rocket uses a 48" main chute there's little value in using a bag with a 36" pilot chute.
In short, in my opinion, consider a bag when the rocket weights around 10 pounds or more. Otherwise the pilot chute can get as big as the main chute.
To ensure that a pilot chute works right (does not invert or tangle) it is often made using nylon mesh rather than suspension lines. Lots of companies make parachutes designed to be used as a pilot chute, but not for rocketry. The typical pilot chute for skydiving costs $100 or more. Spherachutes has made the pilot chutes for rocketry, with nylon mesh, for me, and they have cost far less than $100. I don't recall the exact price. The price for a 36" pilot has been around Spherachute's usual price for a 36" chute.
Why not make the main chute with nylon mesh? Because it gets too heavy and can fail to open. The weight of mesh on a small (36") chute isn't significant. The weight on a large chute becomes significant.
A deployment bag, if made of the right material, such as Nomex, can provide some protection from the ejection charge, but not a lot. A piston or sabot will protect a nylon chute from the heat and flame of the ejection charge.
A deployment bag will not ensure that a parachute exits a rocket. The most commonly used method to extract the parachute from the rocket is to blast the cone off and use it to pull the chute out. Sometimes that works. Sometimes it doesn't work. I prefer to use a piston to shove the chute out, and protect the chute. (You don't want to melt many 10'+ chutes. It gets expensive.) My piston design is not the same as PML's piston design. Piston design is a separate topic.
Summary: In my opinion, for most rockets that weigh less than 10 pounds, use Nomex/Kevlar heat shields between the parachute and ejection charge. You might want to use two.
Dean