For more than five decades, the shape and interactions of the heliosphere with the local interstellar medium have been discussed in the context of two competing models, posited in 1961 1 : a magnetosphere-like heliotail and a more symmetric bubble shape. Although past models broadly assumed the magnetosphere-like concept, the accurate heliospheric configuration remained largely undetermined due to lack of measurements. In recent years, however, Voyagers 1 and 2 (V1 and V2) crossed the termination shock -the boundary where the solar wind drops -north and south of the ecliptic plane at 94 au 2,3 and 84 au 4 in 2004 and 2007, respectively, and discovered the reservoir of ions and electrons that constitute the heliosheath, while Cassini remotely imaged the heliosphere 5 for the first time in 2003. Here we report 5.2-55 keV energetic neutral atom (ENA) global images of the heliosphere obtained with the Cassini/Ion and Neutral Camera (INCA). We compare them with 28-53 keV ions measured within the heliosheath by the low-energy charged particle (LECP) experiment onboard V1 and V2 over an 11-year period (2003-2014). We show that the heliosheath ions are the source of ENA. These observations also demonstrate that the heliosphere responds promptly, within ~2-3 years, to outward propagating solar wind changes in both the nose and tail directions. These results, together with the V1 measurement of a ~0.5 nT interstellar magnetic field 6 and the enhanced ratio between particle pressure and magnetic pressure in the heliosheath 7 , strongly suggest a diamagnetic bubble-like heliosphere with few substantial tail-like features. Our results are consistent with recent modelling 8-11 .Charge exchange between ions and the neutral hydrogen gas flowing through the heliosheath generates the energetic neutral atoms (ENA) that are sensed remotely by Cassini/INCA (see Methods), and enable images of the celestial sphere 12 that place the local measurements by each Voyager in a global context (see Methods). A conceptual representation of our ENA observations in a three-dimensional format, together with key heliospheric boundaries identified by the two Voyagers, is illustrated in Fig. 1a. Unlike the tail-like heliosphere model (Fig. 1b) adopted following Parker's 1961 1 calculations using a subsonic, incompressible hydrodynamic stellar wind flow (referred to henceforth as 'Parker 1'), Fig. 1a does not include an extension of the heliosphere in the heliotail ('anti-nose') direction. This resultthat is, the rough 'tail to anti-nose' symmetry in Fig. 1a -has been made possible by the correlation of 11 years of INCA ENA images and in situ V1/V2 ion measurements with the long-term variability of solar cycles 23 and 24. We infer that the heliosphere is consistent with Parker's 1961 'alternative' notion (henceforth 'Parker 2'), which presents a bubble-like structure formed under the influence of a large-scale interstellar magnetic field (depicted by the grey lines in Fig. 1a) that confines the heliosheath plasma nearly symmetrically in all directions...