Academia should be faster than it is. This especially applies to the transportation and planning journals with which I am familiar. It often takes more than a year to do less than 8 hours of work (reviews and editorial decision-making).
Some peer reviewed journals (which will remain nameless in this post) are best described as Black Holes. Articles are submitted to be reviewed and never escape. There may or may not be an acknowledgment of receipt. The paper may or may not have been sent to reviewers. The reviewers may or may not have acknowledged receiving the paper in a timely fashion. They may or may not conduct the review in a timely fashion. Some reviewers might do their job, but the editor may be waiting on the slow reviewer before making a decision.
There are several causes for this black hole:

1. Authors – Why would you be foolish enough to place your trust in an editor you don’t know? But of course, for graduate students and tenure track faculty, what choice do you have when your career is determined by success in the publication game? In this game, the author is in general the supplicant. If the author is famous, the situation might be reversed, and the editor should be seeking your paper to make their journal stronger, but given there are well over 1 million peer-reviewed journal articles published each year (and I guess 2x that submitted), and only 20,000 Scopus indexed journals, the average journal has leverage over the average author.
2. Reviewers – Why would I do free labor for a stranger (the Editor) for a community I don’t know (prospective readers) to help an anonymous person (the Author)? Why would I do it quickly?
   • The noble answer is to stay on top of cutting edge research.
   • A plausible answer is the opportunity to ensure your own work is properly referenced. Though this might appear sketchy to you as an author when reviewers say cite X, Y, and Z (and reveal themselves), yet you as an author will still cite these works in the revised manuscript, and it looks perfectly natural to the reader. This motivates the reviewer and raises the citation rate of the reviewer’s own works.
   • Another answer is to accumulate social capital.
     Where exactly do I redeem the social capital I am accumulating? Where is the social capital bank:
     Editors write promotion (or immigration!) letters in support of good, quick, helpful reviewers. Editors might more favorably view the papers of helpful reviewers. Editors might more favorably review proposals of helpful reviewers. Editors might be more likely to nominate good reviewers for awards. Editors might nominate good reviewers to an Editorial Advisory Board and bestow upon them some prestige. The reviewer might be an editor elsewhere and be able to “return the favor”. But all of this is probabilistic and a bit vaporous. Journals sometimes publish list of reviewers. In any case, a list of (self-reported) number of reviews by journal is one of the beans that is counted in the promotion and tenure process.
3. Editors – What leverage do I have over unpaid labor (reviewers) and why should I care personally about ungrateful authors who have submitted an unready paper to my journal which will almost inevitably not be accepted the first round. The leverage is future favors I might bestow in advancement of potential reviewers (see above). This indicates Editors should favor graduate students and assistant professors as reviewers over full professors. Unfortunately full professors are more famous and more likely to be selected to do reviews. I am personally running at a rate of about 100 review requests per year now. If I were really famous, I would need to decline far more than I do now. If I were really, really famous, I would not have time to decline requests (or perhaps I would have staff decline requests for me).

So there is a social network at play in this process, and if any link breaks between author and editor, between editor and reviewer, or back from reviewer to editor or editor to author, the circuit is not complete, the paper entered the system, and like a light from a black hole, cannot escape.
This is one reason I like journals that have check to automatically track publication status, nag reviewers, and have quick turnaround times. This is one reason I like the idea of “desk reject”. It is much better to be rejected immediately then after 6, 9, 18, 24 months of review. Fast has value.
There is a second black hole, not quite as large, dealing with accepted papers that have yet to be formatted for publication. This is usually solved by an online “articles in press” or “online first” section of the journal website. The advantage to the journal of this is the ability for papers to accumulate citations before they are actually “published”, thereby gaming the ISI impact factors, which look at the number of citations in the first 2 years from publication.
A major problem with looking at 2 years when journals are slow is apparent. I cite only papers published before I submit my paper. If it takes 2 years to accept and publish, I will not have included any papers from the past two years. Therefore slow fields have lower impact factors than fast fields. This feeds the notion (in a positive feedback way) that these fields are sleepy backwaters of scientific research rather than cutting edge fields where people care about progress.
To break the black holes I have a couple of ideas:

1. A “name and shame” open database (or even a wiki) which tracks article submissions by journal, so that authors have a realistic assessment of review, and possibly re-review and publication times. Also the amount of time in the author’s hands for revision would be tracked.
2. Money to pay reviewers and editors to act in a timely fashion and publication charges to finance open scholarly communication. A few journals pay reviewers. When I get one of those, I am far more likely to review quickly than when I get requests from other journals, especially for journals outside my core area, especially when the likelihood of withdrawing social capital is minimal. Other journals charge authors and use the funds to speed the process (but as far as I know these journals don’t pay reviewers). Of course, we need to be clear to avoid “pay to play”. Libraries could help here, redirecting funds from the traditional subscription model to a new open access model, helping their university’s authors publish in truly open access journals. The new federal initiative will hopefully tip the balance.

We all know the journal system as we have known it is unlike to survive as is for the next 100 years. It is surprising it is lasted as long as it has, but academia is one of the last guilds.
There are lots of cool models out there beyond the traditional library pays for subscription of expensive journal: from open access journals with sponsors (JTLU), author fees (PLOS_One), membership (PeerJ), decentralized archives (RePEc), and centralized electronic archives arXiv.
Yet we need some way of separating the wheat from the chaff, and peer-review, as imperfect as it is, has advantages over the open internet where any crank can write a blog post.
Eventually time will act as a filter, but peer-review, the review of papers by experts to filter out the poorly written, the wrong, the repetitive and the redundant, can save readers much time.