This is a paragraph from an abstract of a paper published in a recent top-tier system conference, anonymized:
In this paper, we characterize the challenges of <the problem we are attacking>. We then introduce <a new abstraction> which exploits the internal <characteristics> to address these challenges. Leveraging <our new abstraction> we introduce a new approach to <attack the problem>. We provide a detailed analysis and experimental evaluation comparing <our new abstraction> to two popular <existing systems>. Finally, we describe three different implementation strategies for <our new abstraction> and discuss their relative merits with empirical evaluations on <real world problems> demonstrating order of magnitude gains.
No numbers. No key-points. Easy template. Little information. Hate it. It would be really better if it is more specific.