Petraeus and his co-authors discussed this strategy at great length in the Army's counterinsurgency field manual. One point they made is that it requires a lot of manpower -- at minimum, 20 combat troops for every 1,000 people in the area's population. Baghdad has about 6 million people; so clearing, holding, and building it will require about 120,000 combat troops.And all you need to know about why it's happening:
Right now, the United States has about 70,000 combat troops in all of Iraq (another 60,000 or so are support troops or headquarters personnel). Even an extra 20,000 would leave the force well short of the minimum required—and that's with every soldier and Marine in Iraq moved to Baghdad. Iraqi security forces would have to make up the deficit.
One widespread, and plausible, theory is that the surge constitutes a last-ditch effort at success. The thinking goes like this: Maybe this will work; and if it doesn't work, the United States can cut its losses and pull back without making the retreat seem like too disastrous a debacle. "We gave it our all," the president could say; "don't blame us that it fell apart." And, since Kagan and other surge-advocates are saying the plan would take about two years to succeed or fail, the next president—not Bush—would be the one who orders, and takes all the heat for, the retreat.
I am not one who likens the Iraq war to Vietnam, but there is an eerie parallel to a memo that John McNaughton, Secretary of Defense Robert McNamara's closest aide, sent to him on March 24, 1965, after it seemed clear that the Rolling Thunder bombing campaign was producing scant results. "The situation in Vietnam is bad and deteriorating," McNaughton wrote. The important aim now is to "avoid a humiliating U.S. defeat (to our reputation as a guarantor)." Therefore, it is essential "that the U.S. emerge as a 'good doctor.' We must have kept promises, been tough, taken risks, gotten bloodied, and hurt the enemy very badly."