We prove that the integrality gap of the Goemans-Linial semide nite programming relaxation for the Sparsest Cut Problem is Ω( log n) on inputs with n vertices, thus matching the upper bound (log n) 1 2 +o(1) of [3] up to lower-order factors. This statement is a consequence of the following new isoperimetric-type inequality. Consider the 8-regular graph whose vertex set is the 5-dimensional integer grid Z 5 and where each vertex (a, b, c, d, e) ∈ Z 5 is connected to the 8 vertices (a ± 1, b, c, d, e), (a, b ± 1, c, d, e), (a, b, c ± 1, d, e ± a), (a, b, c, d ± 1, e ± b). This graph is known as the Cayley graph of the 5-dimensional discrete Heisenberg group. Given Ω ⊆ Z 5 , denote the size of its edge boundary in this graph (a.k.a.We show that every subset Ω ⊆ Z 5 satis es |∂ v Ω| = O(|∂ h Ω|). This vertical-versushorizontal isoperimetric inequality yields the above-stated integrality gap for Sparsest Cut and answers several geometric and analytic questions of independent interest.The theorem stated above is the culmination of a program that was pursued in the works [7, 19-22, 53, 56] whose aim is to understand the performance of the Goemans-Linial semide nite program through the embeddability properties of Heisenberg groups. These investigations have mathematical signi cance even beyond their established relevance to approximation algorithms and combinatorial optimization. In particular they contribute to a range of mathematical disciplines including functional analysis, geometric group theory, harmonic analysis, sub-Riemannian geometry, geometric measure theory, ergodic theory, group representations, * The research presented here was conducted under the auspices of the Simons Algorithms and Geometry (A&G) Think Tank. A full version of this extended abstract, titled "Vertical perimeter versus horizontal perimeter," that contains complete proofs and additional results is available at https://arxiv.org/abs/1701.00620. Nevertheless, this extended abstract contains material that is not included in the full version.