A Metasearch engine is simply search-engine software which uses other search-engines output as a basis for it's own results. It takes search-queries as input from users and instantly sends these out to multiple sources. The results are then parsed, sorted and presented as a list of results.
Advantages and disadvantages
One clear and obvious advantage of using a metasearch engine over a regular one is the diversity of results. If two of three search-engines is censoring a site or topic it's fine, the results from the third uncensored engine will fill in the blanks.
Speed is an issue which can not really be solved; a metasearch engine will have to pull results and sort them to function. This will by nature always be slower than regular search-engines.
Unpopularity among search-engines and outright banning is a big problem with public instances for smaller sites running free software metasearch engines. It is specially problematic for popular ones. If you run a real search-engine and you look at your traffic then you can plainly see that this IP over there is sending 2000 queries per minute. It is blatantly obvious that's not traffic from some average person using a web browser. This has a clear disadvantage for operators and users of meta-search engines: A publicly listed Searx instance, for example, may work well one day and the next day there's barely no results because two major search-engines have stopped providing it with results. Commercially operated metasearch engines avoid this by paying a fee which allows them to use special private APIs.
How common are metasearch engines?
It's actually a lot more common than you'd think. Many "search-engines" such as Yahoo, Yippi, Ecosia ask Bing to produce their results. Startpage gets it's results from Google. DuckDuckGo and EntireWeb queries Bing and Yandex.
Metasearch Software you can Install and Use
Searx is the most functional and modern metasearch engine for Linux desktops and servers alike available today. It's written in Python. You can easily install it on your own machine or use one of the public instances on the Internet. You can read more about it in it's own article.